Sie sind auf Seite 1von 92

Landmark 2003 OpenWorks Troubleshooting Guide 排错指南(学习笔记)

http://www.rainlane.com/dispbbs.asp?BoardID=11&replyID=48712&id=16555&skin=0

Contents 目录
Overview 简介
What's in this Manual

Basic Troubleshooting 排错初步


What's in this Chapter
Symptoms of Trouble
Troubleshooting Steps
Checking Essential System Processes
Checking Which Processes Are Running
Starting Processes
Killing Processes
Telneting from Another System and Using the Kill Command
Killing and Restarting X
Checking Disk Space
Last Resort
Where to Look for More Help
Other Useful Information

Tools for Troubleshooting 排错工具


What's in this Chapter
System Resource Monitor
Types of Resources Monitored
Optimizing the SRM on a Server
Running the Motif Version of the SRM
Running the ASCII Version of the Monitor
Troubleshooting the SRM
Database Sanity Checker
Starting and Stopping the Sanity Checker
Using the Sanity Checker Main Window
Using the ORACLE Screen
Environment Status Tool
Checking the OpenWorks Environment
Project Query
Using LGC_DEBUG for SeisWorks Problems
Activating/Deactivating LGC_DEBUG
Using LGC_DEBUG
Example of Output
Troubleshooting Data-Related Problems 解决与数据有关的问题
What's in this Chapter
Data Troubleshooting Checklist
SeisWorks Users: Do You Have an OpenWorks Project?
All Users: Is Your OpenWorks Environment Properly Configured?
All Users: Can You Access ORACLE?
ORACLE Environment Variables
Troubleshooting Data Access
Data Access Problems in OpenWorks
Data Access Problems in SeisWorks
Problems with Horizons inSeisWorks
Cannot Make Another Horizon
Horizon List Incorrect orMissing
Displaying Horizons in Map View
Problems with Faults in SeisWorks
Correlating Faults
Identifying Faults in Map View
Suspect-Looking Fault Contours
Problems Exporting Data from SeisWorks
Checking the Continuity of SeisWorks/2D Data Files
Data Loading Problems
General Data Loading Problems
SeisWorks Data

Troubleshooting Seismic Display Problems 解决地震显示问题


What's in this Chapter
Problems with Seismic Display
No Seismic
Snow or Stripes on Display
Gaps or White Spots on Seismic Line
Problems with Colors
Can't Get Dual Color Bars
No Data Displayed
No Color Maps
Partial Data Displayed
Other Display Problems in SeisWorks
Troubleshooting Computed Contours
Troubleshooting Your Well Display

Troubleshooting Hardcopy Problems 解决硬拷贝问题


What's in this Chapter
Initial Checklist
Resetting Plotex
Administrator's Console or Operator Console Assisted Purge
Manual Purge
Troubleshooting the Help System 解决帮助问题
Introduction
What's in this Chapter
Help System Error Messages
Help Menu Does Not Work
Help Documents Disappear
Troubleshooting System Problems
What's in this Chapter

Checking OpenWorks System Processes 检查OpenWorks系统进程


Starting/Stopping System Processes
Processes Initiated on Startup
Modifications to System Files
Troubleshooting System Problems
Troubleshooting X
Problems Starting SeisWorks

Troubleshooting ORACLE Problems 解决ORACLE 问题


Introduction
What's in this Chapter
Before Using These Procedures
Documenting and Reading Error Messages
OpenWorks Requirements for ORACLE
Recommended Settings
Recommended Datafile Sizes
Database Block Size
ORACLE Parameters
Applying the Changes
The Environment Status Tool and ORACLE
Cannot open the database instance configuration file
Using Server Manager to Query the ORACLE Database
ORACLE in a Networked Environment
The listener.ora file
The tnsnames.ora file
Troubleshooting ORACLE Problems
Database Processes May Not Be Running
Unable to Connect to a Local Database
Unable to Connect to a Remote Database
Unable to Halt the ORACLE Processes
Cannot Shut Down Database
Cannot Start Database
Object or Data Lost
Database Resource Limits
Database Connection Problems

Common Error Messages 常见错误信息


Introduction
What's in this Chapter
Setting the Error Level
Checking the Error Level
Changing the Error Level
Viewing the Error Messages
Using the OpenWorks Error Logger
Common OpenWorks Error Messages
“The master database owsys could not be connected”
“Error reading the OpenWorks system database (owsys)!”
“Unable to connect to the default $ORACLE_SID database”
“Cannot open the database instance configuration file”
“OpenWorks Database Project *nowell* Initialized”
“OpenWorks Database Project ‘’: Access Failed”
“You must be the Project Owner to Use this.”
Common SeisWorks Error Messages
“Cannot open .sm files. Cannot continue.” (SeisWorks/2D)
“Cannot set project” or “Cannot open project definition file”
“Error in survey parameters in “.pdf” and “.hrz”
“No fault segment/heave selected for assignment.”
“.pd2 file not present” (SeisWorks/2D)
“Viewport Not Allocated”
“Virtual Graphics Error”

Requesting Technical Assistance 寻求技术支持


Introduction
What's in this Chapter
Opening the TAR Tool
Using the TAR Tool
Filling Out the TAR Form
Opening a New Form
Saving a Form
Printing a Form
Sending the Form by E-Mail
Setting Defaults for the TAR Tool
Setting Default Company and Customer Service Names
Setting a Default Printer
Changing the E-Mail Address
Applying the New Defaults to the TAR Too
Exiting the TAR Tool

Index 索引

What's in this Manual


This manual contains the following chapters:
• Basic Troubleshooting. This chapter contains some simple techniques you can use to correct
problems. Refer to this chapter at the first sign of trouble. If the solutions described here do not solve
your problem, refer to one or more of the more detailed chapters in the rest of the manual.

• Tools for Troubleshooting. This section describes the OpenWorks utilities you can use to verify that
system processes are running. These tools can help you pinpoint the causes of problems.

• Troubleshooting Data-Related Problems. This chapter provides solutions to some problems you
might encounter accessing projects or data, as well as problems with importing or exporting data.

• Troubleshooting Seismic Display Problems. This chapter can help you with seismic display
problems, color bar problems, and problems displaying computed contours and well data.

• Troubleshooting Hardcopy Problems. This chapter contains solutions to problems with queuing and
printing hardcopy and plotfiles.

• Troubleshooting the Help System. Use this chapter when you are unable to access help files for an
application.

• Troubleshooting System Problems. This chapter describes the system processes used by Landmark
applications and provides tips for solving various problems with OpenWorks and SeisWorks.

• Troubleshooting ORACLE Problems. This chapter provides detailed steps you can take to correct
problems with your ORACLE database.

• Common Error Messages. This chapter contains the procedure for setting your system’s error level,
which determines the types of error message that will appear. This chapter also describes typical system
error messages.

• Requesting Technical Assistance. This chapter tells you how to use the TAR Tool to report problems
to Landmark support personnel.

Basic Troubleshooting 排错初步


This chapter explains troubleshooting techniques you can use if you encounter problems while using
OpenWorks or another Landmark product. In addition to reading the information in this chapter, you
can call your Landmark customer service representative.

What's in this Chapter 本章内容


This chapter contains the following sections:
• Symptoms of Trouble 故障现象. This section lists several specific symptoms of trouble that may
occur and suggests a procedure for what to do if one of these problems occurs.
• Checking Essential System Processes 检查系统关键进程. This section explains how to check which
UNIX system processes are running, how to restart a stopped process, and how to kill (stop) a process
when the process cannot be stopped by normal means. This section also describes important system
processes and files that you may check when troubleshooting.

• Checking Disk Space 检查磁盘空间. This section explains how to determine the amount of disk
space is available on your system.

• Last Resort 最后的办法. This section tells you how to reboot your system, which you should only
do if the steps described in the preceding steps have failed to resolve your problem.

• Where to Look for More Help 浏览其他章节,寻求更详细的帮助. This section describes the
remaining chapters of this manual, which contain more detailed troubleshooting procedures.

Symptoms of Trouble 故障现象


General symptoms of trouble can include any of the following:

• All processes stop working (system lockup)


• The mouse cursor does not move
• An existing window disappears
• Additional windows cannot be opened
• Individual applications do not respond to user selections
• Command Menu does not respond to user selections
• Command Menu (launcher process) is lost
• The window title bars are lost (mwm process)
• Inter-task communications cannot be performed (pd process)

Methods for dealing with each of these problems are explained on the following pages.

Many problems can be remedied quickly using non-diagnostic strategies. Often you can resolve
problems by exiting the application you are using and then reopening the program. Try this approach if
the
windows are not opening properly, data appears to be inaccurate, faults are posted in the wrong place,
horizons are not properly posted, etc.

Troubleshooting Steps 排错步骤


If closing and reopening the application does not work, or if you are unable to close the application, try
one or more of the following:

• If an application you are using becomes unresponsive, you might need to kill it. See “Killing
Processes” on page 8.

• Expand your Command Menu to show the status of vital processes. Start any processes that are not
running. See “Checking Essential System Processes” on page 5.

• Exit OpenWorks using the Exit command from the OpenWorks command menu, and restart
OpenWorks.

Restarting OpenWorks restarts all individual tasks. If restarting OpenWorks fails to correct the problem,
you can attempt to further diagnose the problem by working with the individual processes (tasks).

Note, however, that in general it is best not to attempt to restart individual applications and tasks.

Checking Essential System Processes


Some UNIX processes are necessary for OpenWorks to run, because they control some functions in
OpenWorks. The information in this section shows you how to check which processes are running,
how to start processes that are not running, and how to kill processes that are running incorrectly.

Checking Which Processes Are Running


OpenWorks provides you with two tools to check on the status of processes that are necessary to run
OpenWorks — a Command Menu that expands to show status and the Environment Status Tool (see
the
chapter entitled Tools for Troubleshooting). The expanding Command Menu shows the status of the
most vital processes.

To check the status of OpenWorks processes, click on the Landmark logo (命令菜单左上角的兰马徽
标 ), located on the left side of the OpenWorks Command Menu. The Runtime Status Box appears
under the command menu.

The most critical OpenWorks processes are listed, preceded by a check box(复选框). If a check box is
green, the corresponding process is running on your system or is available from a server. If the box is
red, the process has not been started on your system or is not available from a server.On CDE desktops,
a check displayed in a red or green check box indicates that OpenWorks has performed a verification on
the status of this resource.

To toggle the status box off, click on the Landmark logo again. To reinvoke the status box (after
pinpointing your problem and correcting it), double click on the Landmark logo. This toggles the status
box off and then back on.

Red and green check boxes can signify表示 different conditions on servers than on clients. The table on
the following page lists these differences.

OpenWorks Runtime Status Box

Process 进程 Server 服务器 Client 客户机


Pointing Dispatcher Green. The pointing dispatcher Green. The specified pointing
specified for your environment dispatcher is running as specified
is running. Red. The pointing for your environment. It may be
local and it may be remote. Red.
dispatcher specified for your
The pointing dispatcher may be
environment is not running. dead. Or it may be inaccessible.

Oracle Green. ORACLE is running. Red. Client cannot communicate


Red. ORACLE is not running. with ORACLE on the server.
Green. ORACLE is running on the
server.

LAM (License Green. Your system is able to check Green. Your system is able to check
Application Manager) out licenses via the specified license out licenses via the specified license
server. Red. Your system is not able server. Red. Your system is not able
to check out a license through the to check out a license through the
specified license server. The license specified license server. The license
server may be dead. Or, it may be server may be dead. Or, it may be
inaccessible. inaccessible.

FrameViewer Green. The FrameViewer is Green. The FrameViewer is


activated. activated.

Red. The FrameViewer is not Red. The FrameViewer is not


activated. activated.

System Resource Green. xsrm is running. Red. xsrm Green. xsrm is running. Red. xsrm
Monitor (xsrm) is not running. is not running.

Remote Services Green. RSvD is available on the Green. RSvD is available on the
Daemon (RSvD) ORACLE server. Red. RSvD is not ORACLE server. Red. RSvD is not
available on the ORACLE server. available on the ORACLE server.

Network Services Green. NetD is available on the Green. NetD is available on the
Daemon (NetD) ORACLE server. Red. NetD is not ORACLE server. Red. NetD is
available on the ORACLE server. not available on the ORACLE
server.

Starting Processes

Restarting the Database


The following procedure will not work if the ORACLE database processes are not running. If your
expanded Command Menu shows that the database is not running, you will have to start the database
processes as explained in “Starting/Stopping System Processes” on page 103.

If a process is not running, you can restart it by typing the following from the UNIX shell window:

processname 2>/dev/null &


For example, if Motif Window Manager stops operating, as evidenced by the loss the borders around your
windows, restart that task by typing (from the UNIX shell window):
mwm 2>/dev/null &

This command tells the system to redirect the standard error stream to /dev/null. The ampersand (&) tells
the system to run this process in the background. This is done so the process will not lock up the window
from which it was started.
The process names are as follows:
Process Name Description Permissions
required

pd Pointing Dispatcher -rwxrwx--x

lam License Application Manager

viewer FrameViewer

mwm Motif Window Manager -rwxrwx--x

Killing Processes

When restarting some individual OpenWorks tasks, you sometimes must kill and restart
other tasks as well for OpenWorks to operate properly. The following chart shows the
tasks you must kill and restart if you restart certain tasks.

If you restart this: You must also kill and restart this:

pd All OpenWorks tasks

mwm None

All other tasks can be started without affecting other tasks.


Sometimes it is necessary to kill or to stop a process that cannot be stopped by normal
means (such as pressing Ctrl-D, or using the Close command from the Windows pull down
selection). If this happens, check the process table for the process in question. (See
“Checking
Which Processes Are Running” on page 5.)

Try This First . . .

Before you attempt to kill individual processes, try to exit OpenWorks via the Command Menu. If you are
able to do this, use startow to restart OpenWorks from the Console Window.
Killing a Process

If you are unable to exit and restart OpenWorks, try to kill one or more processes. To stop a
process, use one of the commands described below.

• To stop pd, type:

stop_pd

• To kill or stop any other current processes, you must first find out the process ID number. To do
this type

ps -ef | grep <processname>

A listing appears showing the process ID. To kill the process, type

kill -9 ####

#### is the process ID number.

• To kill multiple processes, type the process numbers separated by spaces. For example:

kill -9 #### $$$$ %%%%

####, $$$$, %%%% are three different process numbers. If you attempt to kill a process that is not
running, this message appears

kill:####:no such process

Telneting from Another System and Using the Kill Command(从其他机器登录


并使用Kill命令杀进程)

If an application hangs and you lose control of the mouse cursor, try killing the process from another
workstation on the system.

1. At another workstation on the system, type

telnet <your_machine_name>

2. When prompted for a login, enter the login you normally use on your machine. The prompt should now
show your machine’s name.

3. Determine the process ID as described above.

ps -ef

4. To kill the application, type


kill -9 <PID#>

5. To exit from the telnet mode, type

exit

The prompt should show the name of the remote machine.

Killing and Restarting X

If you can’t kill the application, you may want to try restarting the X server.

On a Solaris System

1. To determine the process ID number for the X server, type

ps -ef|grep mwm

ps -ef|grep .xinitrc.r5

2. To kill X, type

kill <PID#>

Do this for both processes found in step 1. This should return you to a white screen.

3. To restart X windows, type

startserver

You may find you need to log out by typing exit, then login with your user name, and then use the
startserver command.

On an IBM System

1. To determine the process ID number for the X server, type

ps -ef|grep X

2. To kill X, type

kill <PID#>

This should return you to a white screen.

3. To restart X windows, type

startserver
On a Silicon Graphics System

On an SGI system, log out and log back in to restart the X server.

1. From the Desktop menu, select the Log Out option.

2. When asked to confirm the logout, select Yes.

3. When the login dialog box appears, select your login ID and click on Log In.

Checking Disk Space


To list all of the current hard drive devices and their directory name equivalents with the total, free, and
percent-used space for each, type from any UNIX shell:

df -k (Solaris and SGI)

df -I (IBM)

All values are shown in blocks where a block equals 1024K bytes.

Checking disk space may provide you with information about where you have space problems.

Last Resort
If all troubleshooting efforts have failed, reboot your system.

Shutting Down the System

If you cannot kill the application from another machine, you may want to shut down the system and then
reboot it. Shutting down the system is safer than performing a panic halt because it forces your file changes
to be written to disk and it closes the open files on your system, preventing them from becoming corrupted.

1. In a remote or local xterm, type su to become superuser.

2. To shutdown your system, type

shutdown <your_machine_name> -h now

A message stating that the system is about to be shut down will be displayed on terminals of users that are
logged into your machine.

The message will also be posted on the terminals of systems that are mounted to your machine.

Moments after the message is posted, your system will be shut down.

3. Once your system is shut down, type boot to reboot it.


For more information on the shutdown command, consult the “shutdown” man page.
(我通常使用这个命令:# shutdown -y -g0 -is)

Always shut down ORACLE before shutting down your system.


Shutting down your system while ORACLE is running can corrupt the
database. Always perform an orderly shutdown of ORACLE before shutting
down your system.

Aborting and Rebooting

If your system hangs and you cannot log into it remotely, you may have to abort and reboot. Aborting and
rebooting is a risky way to restart your system because it does not force your file changes to be written to
disk and it does not close the open files on your system. As a result, if you abort and reboot, you risk losing
changes and corrupting open files.

This is a particular danger for an ORACLE server.

Before you abort, make sure that the system is truly dead. It is possible that you are experiencing a network
problem, which could cause delays in system activity. If this is the case, wait until the network server
responds.

The following instructions explain how to abort and reboot.

On a Sun system

1. Hold down the Stop key and press a.

You will receive a message prompting you to reboot the system.

2. In response to the message, press b for boot.

On an IBM system

Press the reset button on the tower.

On a Silicon Graphics system

Press the reset button. When rebooting is completed, restart OpenWorks and SeisWorks

Do not abort and reboot unless instructed to by a Landmark


customer support representative.
Shutting down your system while ORACLE is running can corrupt the
database. You may have to restore your database from the most recent backup.
In this case, you will lose the changes that you have made in the interval.

Where to Look for More Help


The preceding pages describe some basic steps you can take to resolve problems with your system. Some
problems, however, are more complex and require more intensive action. The chart below describes various
types of problems and their symptoms and refers you to the appropriate section of the manual to help you in
resolving problems.
Type of Problem Symptoms Refer to

•trouble accessing projects or data


Data Access “Troubleshooting Data-Related
Problems” on page 49
•trouble displaying well, curve, or
fault data

•problems with importing or


exporting data

•problems with seismic displays,


Display “Troubleshooting Seismic Display
such as snow or gaps in seismic lines
Problems” on page 81

• problems with color display

•trouble with computed contours or


well displays

Hardcopy •problems sending files to a plotter “Troubleshooting Hardcopy Prob-lems”


on page 87
•trouble generating printed out-put
from a plotter

Help System •inability to display help docu-ments “Troubleshooting the Help Sys-tem” on
for an application page 97

•system processes not running


General System Problems “Troubleshooting System Prob-lems” on
page 101
•trouble with the X Window system

•problems starting SeisWorks

ORACLE •ORACLE processes not running “Troubleshooting ORACLE Prob-lems”


on page 109
•inability to access the ORACLE
database

Other Useful Information

Three other chapters in this manual can be useful when you are attempting to solve system problems.

Chapter Description

“Tools for Troubleshooting” on page 17 Describes the OpenWorks utilities you can use to verify that
system processes are running. These tools can help you
pinpoint the causes of problems
“Common Error Messages” on page 131 Contains the procedure for setting your system’s level of
error message reporting and describes typical system error
messages.

“Requesting Technical Assistance” on page 145 Tells you how to use the TAR Tool to report problems to
Landmark support personnel.

Tools for Troubleshooting


What’s in this Chapter
OpenWorks provides several tools to help you check the status of your Oracle database and to troubleshoot
problems in your OpenWorks environment, as follows:

• Environment Status Tool on page 18. Checks if the environment is correctly configured for OpenWorks
and the ORACLE system.

• Project Query on page 20. Uses SQL to query the current Oracle database on the size, location, and user
accessibility of a particular OpenWorks project.

• Using LGC_DEBUG for SeisWorks Problems on page 22.

Describes the LGC_DEBUG utility, which allows you to produce a report of files accessed by SeisWorks.
This report can help you determine causes of data access problems.

Environment Status Tool


The Environment Status Tool lets you check if your environment is correctly configured for OpenWorks
and underlying RDBMS. It provides information about key elements of the runtime environment, system
resources, and system configuration.

The Environment Status Tool assumes there are no RDBMS-related problems. For this reason, before you
invoke the Environment Status tool, you should confirm that Oracle is installed and functioning correctly.

To open the Environment Status tool from the OpenWorks Command Menu, click on Utilities à
Environment Status Tool. The following window appears.

Checking the OpenWorks Environment

The first thing that you should do when you have a problem displaying your wells is to determine whether
the Oracle database is properly installed and running. Next, you should consult the Environment Status
Tool to determine whether the OpenWorks environment is properly configured.

Select Status à System from the menu bar of the Environment Status Tool. The System Status Menu
appears.

(依此操作:OpenWorks 2003 Command MenuàUtilitiesàStatusà (Oracle,System) 即可查看有关的


环境状态)

Project Query 工区查询


The Project Query tool lets you check the location, size, and user accessibility to database tablespaces for a
project. You do not need to be assigned user access in order to query the database about a particular project.
Project Query will provide information about any of the Oracle projects in the database.

You invoke Project Query as follows.

1. In the OpenWorks Command Menu, click on Project àProject Admin. The Project Administration
dialog box appears.

2. In the Project Administration dialog box menu bar, select the project of interest and then click on Project
àQuery. You will receive a window similar to the one shown below.

Project Query has a number of read-only information fields that provide details about the size, space, and
accessibility of the Oracle project. Many of the entries in this dialog box refer to the project “tablespace.”
This is the disk space allotted to the project by Oracle.

The following table explains the meaning of each field:

Project Query Fields

Field Name Description

Project Availability Project status. Currently, the only status available for this field is
ONLINE.
Project Default Table Extent The tablespace consists of one or more files. To store data for a
Parameters particular table within the tablespace, space is allocated in
blocks called a “extents.” When a table is created, it is
allocated an initial extent. As data is added, if the current extent
is used up, an additional extent is allocated.

A size may be specified for the initial extent and for the second
extent allocated (or the Next Extent). All subsequent extents
will be allocated in the size of the Next Extent. If a Percent
Increase has been specified, all subsequent extent allocations
will be a certain percent larger than the previous one.
Directories, names, and sizes of files that collectively make up
the “tablespace.” You use these names in SQL queries.
When you create an OpenWorks project, you are given the
Tablespace File option of making a small, medium, or large project. These
result in tablespace sizes of approximately the following sizes:

small = 60 MB

medium = 100 MB

large = 200 MB

You can add disk space to the existing tablespace using the
Project Administration: Modify option.
Free Space Available Space available for additional project data in the tablespace.

Number of Tables Number of tables in the tablespace.

The names of users currently granted access to the Oracle


project. Each name listed with the category of access that has
Project Users been granted to the user.

• BROWSE entitles the user to look at the project.

• INTERP entitles the user to create, view, modify, and delete


their own data.

• MANAGE entitles the user to delete, modify, backup, and


restore projects, as well as create, view, modify, and delete
project data. It also allows the user to add, update, and delete
project users.

For more information on Oracle and Oracle troubleshooting, refer to “Troubleshooting Oracle Problems”
on page 71.

Using LGC_DEBUG for SeisWorks Problems


The environmental variable LGC_DEBUG produces a report on files as they are accessed by the SeisWorks
input/output system. This information may assist you in identifying a variety of problems. For example, if it
shows that all attempts to open the project definition file have failed, you know immediately to check the
permissions on and location of that file.

Activating/Deactivating LGC_DEBUG

To activate LGC_DEBUG, you set it and start SeisWorks from the same xterm window. The commands
differ slightly, depending on which type of shell you are using.

In a Bourne Shell

To set LGC_DEBUG, use the following commands:


LGC_DEBUG=1

export LGC_DEBUG

To turn off LGC_DEBUG, type

unset LGC_DEBUG

In a C Shell

To set LGC_DEBUG, use the following commands:

setenv LGC_DEBUG 1

To turn off LGC_DEBUG, type

unsetenv LGC_DEBUG

Using LGC_DEBUG

To use LGC_DEBUG to diagnose problems, perform the following steps.

1. In an xterm, set LGC_DEBUG to 1, using the appropriate script for your system’s shell. (See above.)

2. Start SeisWorks from that same xterm by typing

s2d (SeisWorks/2D)

or

s3d (SeisWorks/3D)

3. When you encounter some problem in running SeisWorks, check the input/output report in the xterm
window. Perhaps the program is unable to open a needed file.

Example of Output(在此省略图示)

An example of the kind of output LGC_DEBUG produces when you start SeisWorks/3D is shown on the
following page. Note that the program searches all project directories for each file (/pa, /pb, /pc, etc.).

So you see many failed attempts to open the file before the program tries the right directory. This is normal.
The problem occurs when all attempts to open a particular file fail.
Beginning of LGC_DEBUG Output for a Successful SeisWorks/3D Session

Troubleshooting Data-Related Problems


Well, curve, seismic navigation, and fault data are stored in OpenWorks projects. If you are having trouble
displaying wells or faults, you need to check that the database is working and that your system is properly
configured to run it.

What’s in this Chapter


This chapter provides solutions for problems related to accessing projects and project data. This chapter
contains the following sections:

• Data Troubleshooting Checklist on page 26. This section contains steps you can take to check for
problems in your OpenWorks or Oracle setup.

• Oracle Environment Variables on page 35. This section describes the environment variables that must
be set for Oracle to function properly.

• Troubleshooting Data Access on page 36. This section contains solutions for common data access
problems in OpenWorks and SeisWorks.

• Problems with Horizons in SeisWorks on page 47. This section discusses problems associated with
creating and manipulating horizons in SeisWorks

• Problems with Faults in SeisWorks on page 49. This section describes solutions for fault data problems
in SeisWorks.

• Problems Exporting Data from SeisWorks on page 51. This section discusses problems you might
have exporting SeisWorks data to Z-Map Plus.

• Checking the Continuity of SeisWorks/2D Data Files on page 52. This section contains the procedure
for using the contest2d utility to check the continuity of 2D master projects.

• Data Loading Problems on page 54. This section provides solutions to problems that can occur after
loading data into a project.

Data Troubleshooting Checklist


Follow the procedures in this section to check for common problems in the OpenWorks and Oracle setup.

SeisWorks Users: Do You Have an OpenWorks Project?

Verify that your seismic project is linked to an OpenWorks project. This association is recorded in the plist.
dat file, which typically is in $OWHOME/conf.

1. In an xterm window, type

plist

You will receive a list of all available Seisworks projects and their corresponding OpenWorks projects. The
SeisWorks projects are listed in the left column. The OpenWorks projects are listed in the right column.
单位机器上的例子:

blade{owuser:/home/blade/owuser}% plist

Project name Creation date


------------ -------------------
db2d 01182006 18:06:54 DB1/ORACLE db1_master
hjq072d 03022007 15:56:24 HJQ/ORACLE hjq_master
hjq2d 04172004 15:02:55 HJQ/ORACLE hjq_master
hjqnew2d 04042006 14:42:02 HJQ/ORACLE hjq_master

blade{owuser:/home/blade/owuser}%

2. Confirm that an OpenWorks project is paired with your seismic project. If none is, use Seismic Project
Associate (available from the Seismic Project Manager’s Project menu) to associate an OpenWorks project
with the seismic project.

SeisWorks/2D users can find more information in the “Master Projects” and “Working Projects” chapters
of the SeisWorks/2D Project Management manual. SeisWorks/3D users should refer to the “3D Projects”
chapter of the SeisWorks/3D Project Management manual.

3. If an OpenWorks project is listed for your seismic project, your problem may be that Oracle does not
recognize you as a user for this project. In this case, perform the following steps:

• Use the Project Query option in the Project Administration utility to determine if you are set up as an
Oracle user for the project (see the OpenWorks Data Management manual for more information on Project
Administration).

If Project Administration does not work, you have a problem in your Oracle or OpenWorks setup. Proceed
to “All Users: Is Your OpenWorks Environment Properly Configured?” on page 27.

• If you are not a valid user, get the project manager to grant you user access.

• If you are a valid user, then the problem lies either with your OpenWorks environment or with the Oracle
setup. Proceed to “All Users: Is Your OpenWorks Environment Properly Configured?” below.

All Users: Is Your OpenWorks Environment Properly Configured? 所有用户:OpenWorks 环


境配置正确吗?

After verifying that you have user access to a valid OpenWorks project, you should look for errors in the
way OpenWorks is set up. Here are a few quick steps to check the OpenWorks configuration.

1. First, determine if the location of the OpenWorks runtime directory is properly defined by the
$OWHOME variable. To do this, type the following command:

lgc_getenv OWHOME

The response should be the top directory of your OpenWorks installation.


实例:blade{owuser:/home/blade/owuser}% lgc_getenv OWHOME

/home/OpenWorksblade{owuser:/home/blade/owuser}%

2. If OWHOME is not properly defined, perform the following steps:

• Using vi or another editor, set the value in your initialization file ($HOME/.lgclogin or $HOME/.
lgcprofile).

• Log out.

• Log in.

• Start OpenWorks.

3. The OWSYSSID variable must also be set properly. To check this variable, type the following
command:

lgc_getenv OWSYSSID

The response should be the name of the Oracle instance where the OpenWorks system database is located.
By convention, the name of the Oracle instance is the node name preceded by ow. Example: for the node
nova, the OWSYSSID would be ownova.

4. If necessary, set OWSYSSID to a correct value either by using the Project Status tool or according to the
following steps.

实例:blade{owuser:/home/blade/owuser}% lgc_getenv OWSYSSID

owblade

blade{owuser:/home/blade/owuser}%

• Using vi or another editor, set the value in your Landmark configuration file ($OWHOME/conf/lgcenv.
cf).

• Log out.

• Log in.

• Start OpenWorks.

5. If NetD is not running, you will be able to access Oracle projects, but you will be unable to create,
modify, or delete projects using Project Administration or Project Create. RSvD allows OpenWorks
applications to access services on remote machines.

To check that the Network Services Daemon (NetD) and Remote Services Daemon (RSvD) are
functioning, perform the following steps:
• In the OpenWork Command Menu, click on the Landmark logo to invoke the Runtime Status Box.

• Look for a green check box next to Network Services Daemon and Remote Services Daemon, indicating
that the daemons are operating.

• If the check boxes are red, query whether the daemons are running. Use the following commands (Solaris,
SGI, and IBM):

ps -ef | grep netd

checkRSvD

• If NetD is not running, start it by typing this command:

start_netd

This should also restart RSvD. RSvD can be initiated only by NetD.

实例:

blade{owuser:/home/blade/owuser}% ps -ef |grep netd

root 161 1 0 07:19:59 ? 0:00 /usr/sbin/inetd -s

blade{owuser:/home/blade/owuser}% checkRSvD

/vobs/ow/src/lib/owenv/owi/owinit.c:138 0 OpenWorks Version 2003 Jul 26 2001 17:17:24

blade{owuser:/home/blade/owuser}% start_netd

The OpenWorks Network Services Daemon is ALREADY running

blade{owuser:/home/blade/owuser}%

6. If, after performing all these checks, you still have trouble accessing the Oracle project, proceed to “All
Users: Can You Access Oracle?” below.

NetD is required on database client machines for Dynamic PD.

If your machine is a database client, the Network Services Daemon needs to be


running locally if Dynamic PD is running. NetD must also be configured and
operational on the host machine that is the database server.

All Users: Can You Access Oracle? 所有用户:都能访问Oracle吗?

At this point you know that you have a valid OpenWorks project, but you cannot reach the Oracle database
from within SeisWorks. The next step is to query Oracle directly using SQL (structured query language 结
构化查询语言).

If this attempt fails, you will know that the problem lies in the Oracle setup.

1. First, check that the environment variable OWSYSSID is properly set. Type this command:

lgc_getenv OWSYSSID

The response should be the name of the Oracle instance where the OpenWorks system database is located.
By convention, the name of the Oracle instance is the node name preceded by ow. Example: for the node
nova, the OWSYSSID would be ownova.

2. If necessary, set OWSYSSID to a correct value either by using the Project Status tool or according to the
following steps.

• Using vi or another editor, set the value in your Landmark configuration file ($OWHOME/conf/lgcenv.
cf).

• Log out.

• Log in.

• Start OpenWorks.

3. If OWSYSSID is properly set, use its value to query the Oracle database. Type this command(输入蓝色
部分):

sqlplus owsys/owsys@<OWSYSSID>

You should get an SQL prompt, indicating that you are now in SQL and can view data in the database.(用
quit退出)

4. If your SQL query is not successful, you need to check whether your command reached the correct
sqlplus. Issue the following two commands:

which sqlplus

[system responds with a path]

echo $ORACLE_HOME/bin/sqlplus

[system responds with a path]

• The “which sqlplus” path should point to the same location as does the echo command. If it does not, the
system administrator may need to redefine your $PATH environment variable. Then repeat Step 3 and
continue from there.

• If the two paths are the same, you have a problem with your Oracle setup. Oracle may not be running, or
the SQL Net Listener may not be running. Proceed to Step 9.
5. If your SQL query is successful, but you are having trouble accessing the database under your own user
name, the problem may be that you are not established as a valid Oracle user. Perform the following steps:

• Change directories using the following command:

cd $OWHOME/bin

• Switch users to the oracle administrator with the following command:

su - oracle

• Invoke the Oracle Database User Management tool with the following command

orauser

The Oracle Database User Management tool should appear in the xterm.

6. Choose Option 4, Show Externally Identified Users of the Oracle Database. A list of valid Oracle
users appears.

7. If you are not in the list of Oracle users, press Return and then choose Option 1, Adding a user to the
Oracle Database. Follow the prompts to add yourself as an external user.

8. Now you must become a project user for the OpenWorks project of interest. Use Project Query in the
Project Administration utility to identify a project user with manager status, and get that user to grant you
access to the project.

9. To check the status of the Oracle environment, select System and then Database Sanity Checker from
the OpenWorks Command Menu.

The Database Sanity Checker appears.

10. From the menu bar of the Database Sanity Checker, select Option and then Oracle.

The Oracle Setup dialog box appears, with information about essential Oracle environment variables, files,
and processes.

11. Check that the environment variables are properly set. Use the following guidelines:

• ORACLE_HOME should be set to the top-level directory of your Oracle installation. (The default setting
is $OWHOME/oracle.)

Also, ORACLE_HOME must point to the actual Oracle directory, not to a link. To check this, go to
$OWHOME and issue an “ls -l” command; this listing will show if a link exists.

• ORACLE_SID should be set to the system ID for the database server your system will be accessing.

If either of these Oracle environment variables is not set correctly, set it using the following steps.
• Using vi or another editor, set the value in your initialization file

($HOME/.lgclogin or $HOME/.lgcprofile).

• Log out.

• Log in.

• Start OpenWorks.

12. In the Oracle Setup dialog box, check that the following Oracle processes are running.

• Recovery Process

• Process Monitor

• System Monitor

• Database Writer

• Log Writer

In a client-server configuration, check processes on the

server.

If your machine is a client and you get Oracle from a remote server, the
Database Sanity Checker will show all Oracle processes as not running (as
indicated by a red radio button). This is normal.

To check the Oracle processes, you must invoke the Database Sanity
Checker from the server where Oracle resides. Perform the following steps:

• Telnet to the server.

• Set the DISPLAY variable to display on your machine.

• Start the Database Sanity Checker with the following command:

$OWHOME/bin/xdbsanchk

13. If any of the five Oracle processes are not running, shut down Oracle and restart it. Use following
commands:

su - oracle

dbshut

dbstart
14. If the SQL*Net V2 Listener Process is not running, OpenWorks applications will not function. To start
the listener, use the following commands:

su - oracle

lsnrctl

start

quit

Both Oracle and OpenWorks can function without Archiver being activated.

15. If you cannot access the database with all essential processes running, the problem may be that your
machine is a client to an Oracle server, and the server is not listed in your tnsnames.ora file.

To check this, perform the following steps:

• To learn the name of your Oracle server, type

lgc_getenv OWSYSSID

• Change directories to $ORACLE_HOME/network/admin.

• Look at the tnsnames.ora file for an entry like this:

实例:

blade{owuser:/home/oracle/OraHome1/network/admin}% ls

listener.ora shrept.lst tnsnames.ora

samples/ sqlnet.ora.bak

blade{owuser:/home/oracle/OraHome1/network/admin}% more tnsnames.ora

# TNSNAMES.ORA Configuration File:/home/oracle/OraHome1/network/admin/tnsnames.ora

# Generated by Oracle configuration tools.

OWBLADE =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = blade.com)(PORT = 1521))


)

(CONNECT_DATA =

(SERVICE_NAME = owblade)

EXTPROC_CONNECTION_DATA.COM =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))

(CONNECT_DATA =

(SID = PLSExtProc)

(PRESENTATION = RO)

blade{owuser:/home/oracle/OraHome1/network/admin}%

where owblade is the system ID of the server being accessed.

• If no such entry exists, or if the existing entry is incorrect, see the OpenWorks System Administration
manual set for complete instructions on modifying the tnsnames.ora file.

You have now verified the following:

• The Oracle environmental variables are properly set.

• Oracle is running.

• SQL Net Listener is running.

• You are a valid Oracle user.

• The Oracle processes are running.


• Your system is properly defined in the tnsnames.ora file (if necessary).

16. If you still cannot access the database, the problem may be an incorrect Oracle listener port. To check
this possibility, select Utilities and then Environment Status Tool from the OpenWorks Command Menu.

The Environment Status Tool appears.

17. From the menu bar of the Environment Status Tool, select Status and then Oracle. The Oracle Status
Menu appears.

18. Note the numeric value for Listener Port.

19. Check that this same value is recorded in the file /etc/services.

This file should have an entry like the following:

tnslsnr 1521/tcp # SQL*Net version 2 listener

In this example, “1521” is the number of the Oracle Listener Port as posted in the Environment Status Tool.

20. If /etc/services does not have a “tnslsnr” entry, add one, or contact your system administrator for
assistance.

21. If the Oracle setup checks out and you still cannot access the database, contact your Landmark customer
support representative for assistance.

Oracle Environment Variables Oracle环境变量


The following variables must be set in order for OpenWorks to run the Oracle database successfully. These
variables will be set when you use the standard Landmark installation procedures to install OpenWorks and
Oracle.

Variable 变量 Where Set 位置 Purpose 目的

ORACLE_HOME C shell: $HOME/.lgclogin Defines pathname of the Oracle


directory. The default is
Korn or Bourne shell: $HOME/.lgcprofile $OWHOME/ oracle.

ORACLE_SID C Shell: $HOME/.lgclogin Specifies the system ID of the


default Oracle instance to be
Korn or Bourne shell: $HOME/.lgcprofile used. In most Oracle
installations, ORACLE_SID is
set to the same ID as
OWSYSSID.
OWSYSSID Default: $OWHOME/conf/lgcenv.cf Identifies the system ID of the
host on which the OpenWorks
Also can be set in the Unix initialization files Oracle instance is installed. By
as follows: C Shell: $HOME/.lgclogin convention, you usually specify
OWSYSSID as the name of the
host prefixed by “ow.” For
Korn or Bourne shell: $HOME/.lgcprofile
example, for a host named nova,
OWSYSSID would be specified
as ownova.

OW_DBTYPE (optional) Default: $OWHOME/conf/lgcenv.cf Indicates which database


OpenWorks is installed on. For
Also can be set in the Unix initialization files OpenWorks, this variable is set
as follows: C Shell: $HOME/.lgclogin to ORACLE.

Korn or Bourne shell: $HOME/.lgcprofile

Troubleshooting Data Access 排查数据存取错误


The OpenWorks data access system is a database or table schema called OWSYS. Information about
OpenWorks projects is kept in OWSYS so that multiple users can have access to the same projects. If you
are having problems with the data access system, such as connection failures or failure to find or open a
project, try the following checks.

Data Access Problems in OpenWorks OpenWorks中的数据存取问题

The following data access problems may occur in OpenWorks installations using Oracle. For more detailed
information on Oracle troubleshooting, see “Troubleshooting Oracle Problems” on page 71.

Environment Variables Are Not Set Properly 环境变量设置错误

The data access environment variables must be set correctly. Use the Environment Status Tool to check
them. The OWSYSSID should be the name of the machine where the system database resides. If a variable
is not set correctly, then check the value of the OpenWorks environment variable OWSYSSID.

To get this value type:

lgc_getenv OWSYSSID

If the node name is not the same as the node where OWSYS is found then you need to edit the file
$OWHOME/conf/lgcenv.cf and correct the OWSYSSID designation so that it has the letters ow followed
by the correct node name. For example, if the correct node name is nova:

OWSYSSID ownova

OWSYS Does Not Exist or Cannot Be Found OWSYS不存在或找不到

Check that OWSYS exists on the node specified by the OWSYSSID environment variable. Do this by
logging on to the node where OWSYS should exist and entering the following command:

sqlplus owsys/owsys

The system should make a connection. If it does not, be sure that Oracle is properly installed and running
(see the “Troubleshooting Oracle Problems” on page 71 chapter).

Also check that you have access from a remote node. To do this, log in to a remote node that should have
access to the node where owsys database exists.

sqlplus owsys/owsys@value_of_owsyssid

If this fails to connect, check that Oracle is running properly at each node. If that is the case and you are
still unable to connect, then you should check that the network privilege to connect between the two nodes
is properly set in Oracle.

OWSYS May Not Have Been Properly Created OWSYS创建的不正确

Once you are in OWSYS you should check to see that the OpenWorks tables exist.

sqlplus owsys/owsys

select table_name, tablespace_name from all_tables

where owner = 'OWSYS';

You should receive a list of over 30 tables.

A Project Database Exists in Oracle, But Is Unknown to OpenWorks 在Oracle中有工区数据


库,但OpenWorks不识别

First check that OWSYSSID is set properly. If it is set properly then the next step is to check that the tables
in OWSYS have the proper values.

There are three tables that should contain information describing a project: ow_sys_project,
ow_sys_min_max, and ow_sys_pal. The ow_sys_project and ow_sys_min_max tables should have one
entry for each project. The name for a project must be unique. The ow_sys_pal should have at least one
entry per project. This entry is the project manager with security level set to 3.

sqlplus owsys/owsys

select * from ow_sys_project;

select * from ow_sys_prj_user;

If the project in question has been deleted from OpenWorks but still exists in Oracle, see also “A Project
Fails to Delete Completely” on page 41.

The Data Access System Cannot Find the Data Files 数据存取系统找不到数据文件
The data access system must have two data files located in the $OWHOME/dat/idb directory. These files
are sql.dat.oracle and xref.dat.

If these files are not found, the GDI initialization will fail and return an error on the command line:

PROBLEM PROCESSING DATABASE CROSS REFERENCE ****:

$OWHOME/dat/idb/xref.dat

or

PROBLEM PROCESSING SQL ****: $OWHOME/dat/idb/sql.dat

If the ERR_LEVEL environmental variable is set to 2 or less, then an error statement is produced:

gdiInit.c:145 217600017 gdiInit: idb_init failure

or

gdiRead.c:365 217600003 gdiRead: Bad sql query

Use the Environment Status Tool to find the $OWHOME variable.

Locate and move sql.dat and xref.dat into the $OWHOME/dat/idb directory to correct this problem. Make
certain that these files have read access permission. Check this with the command

ls -l $OWHOME/dat/idb

A screen response should be similar to:

total 489

-rw-r--r-- 1 owadm 198670 Dec 2 20:33 sql.dat.ingres

-rw-r--r-- 1 owadm 198444 Dec 2 20:33 sql.dat.oracle

-rw-r--r-- 1 owadm 73776 Dec 2 20:32 xref.dat

If three r’s do not exist in the second entry for each file, access may be denied to these files. To correct this
situation, on the command line enter

chmod 444 *.dat

A Data Access Failure Occurs 数据存取失败

If relations exist between database verification tables, yet these relations are not specified, then the entire
access may fail. For example, if an attempt is made to read a casing instance without a formation name and
the vc_stratunitname has no “UNKNOWN” entry, then the access will fail.
The verification tables (VC 验证表) have an ID and a name as their basic components. There must be one
entry that has an ID of zero value and a default name of UNKNOWN, so that unspecified relations can
return this instance, allowing the entire access to succeed. For example, the vc_stratunitname table needs an
entry

strat_unit_name_id = 0

strat_unit_name = “UNKNOWN”

This allows successful access of a casing without having to specify the formation name.

A Depth Preference Is Not Set 未设置Depth优先选项

If unexpected depth values are set, the depth preference may not be set. The default is the project depth
units and measured depth.

A Priority List is Not Set 未设置优先级列表

If no priority list is set and an api routine uses the priority list, then an error message will occur:

Priority list has not been set

If the priority list is not alphanumeric, an error message is produced:

Priority list not alphanumeric

If the priority list is too long, an error message is produced:

Priority list has entry too long or without null terminator.

You Are Not the Owner of a Data Type 你不是数据类型的属主

There are certain data types that check for ownership when a modification is being made. These are

• Calculated Lithology

• Calculated Lithology Header

• Horizon Parameters

• Interpreted Drilling Show

• Pick

• Well List

• Well Note Pad

• Zone
• Zone Attribute

• Zone Detail

If an update or deletion is attempted by a user who does not own the instance, then error messages occur if
the ERR_LEVEL is set at 0 or 1.

Not owner, cannot delete.

or

Not owner, cannot update.

Check ownership in Well Data Manager.

Ancillary Messages That Can Be Ignored 通常可能忽略的辅助信息

The following messages can usually be ignored:

117600004: gdiReadWellPick: idb_fetch_structure failed

117600016: gdiAdd completed successfully

The numbers are what are important.

A Project Fails to Delete Completely 工区不能彻底删除

On rare occasions, a project you delete in OpenWorks may fail to completely delete from the Oracle
database, although the name no longer appears in any project list. In that event, you cannot use that project
name again with a project create or restore command.

In this circumstance, the following commands can be issued to manually delete all remnants of the project
from the database. This approach should be used only by the Oracle database administrator and only on any
rare occasion when a normal project delete has not succeeded.

1. Login as the oracle dba account owner (e. g., oracle) and use the Server Manager utility to access the
privileged Oracle user called internal:

login: su - oracle

password: <dba_password>

svrmgrl

connect internal

2. Now issue all commands needed to fully delete the project “PROJECTNAME” from the database (here,
indented lines represent wrapped lines):
alter session set current_schema=owsys;

delete from ow_sys_project where

project_name=’PROJECTNAME’;

delete from ow_sys_prj_delete where

project_name=’PROJECTNAME’;

delete from ow_sys_prj_user where

project_name=’PROJECTNAME’;

delete from ow_sys_prj_archive where

project_name=’PROJECTNAME’;

delete from ow_sys_prj_bkup_log where

project_name=’PROJECTNAME’;

delete from ow_sys_prj_bkup_freq where

project_name=’PROJECTNAME’;

delete from owsysp.ow_sysp_prj_security where

project_name=’PROJECTNAME’;

3. You must also delete all tablespace related to the project. First, locate the proj.dbs file:

connect internal

select * from dba_data_files;

4. Make note of the path to the project data file (i.e., the location of the proj.dbs file).

5. Issue the following commands to complete the deletion and to exit:

drop tablespace PROJECTNAME including contents

cascade constraints;

drop user PROJECTNAME cascade;

drop role manage_PROJECTNAME;

drop role interp_PROJECTNAME;

drop role browse_PROJECTNAME;


exit

6. Finally, change to the directory where the proj.dbs file is located, and remove it.

cd <path_to_proj.dbs>

rm proj.dbs

Data Access Problems in SeisWorks SeisWorks中数据存取问题

When you move interpretations from 2DPlus or 3DPlus to SeisWorks, you could encounter problems when
trying to access the old data files.

This section describes the solution to these problems.

Your Project Does Not Appear 工区没出现

If you press List in the Main Menu window and your project does not appear in the Project Selection dialog
box, perform the following steps:

1. In an xterm window, type this command:

plist

A list of all your SeisWorks projects and their associated OpenWorks and master projects will be displayed.

实例:

blade{owuser:/home/blade/owuser}% plist

Project name Creation date

------------ -------------------

db2d 01182006 18:06:54 DB1/ORACLE db1_master

hjq072d 03022007 15:56:24 HJQ/ORACLE hjq_master

hjq2d 04172004 15:02:55 HJQ/ORACLE hjq_master

hjqnew2d 04042006 14:42:02 HJQ/ORACLE hjq_master

blade{owuser:/home/blade/owuser}%

2. If the missing project is not listed, edit the plist.dat file. (This is the file from which the list of available
projects is generated.)

• Change directories to $OWHOME/conf


• Use vi to edit plist.dat and add the project.

3. Check that the project directories exist and have open permissions.

• If the problem is permissions, use the following command to recursively give open permissions to each
directory and its subdirectories. You must have superuser status when you issue this command:

chmod -R 777 <directory_name>

• If the project directories are missing, reload them from a backup tape.

4. Check that the project definition file (.ps2 in SeisWorks/2D or .pds in SeisWorks/3D) is in the project
sys directory and has open permissions.

Problems with 2DPlus Files 2DPlus文件有问题

In 2DPlus every data file you created was owned by lgc and had a user ID of 200 and a group ID of 10.
Now, in SeisWorks/2D you have the latitude to establish different user and group IDs.

If SeisWorks users are using different accounts besides lgc with different user and group IDs, they may run
into permission problems when they try to access old 2DPlus files. Simply change the permissions to grant
open access as described in “Changing the

Permission Mode of a File” below.

Problems with 3DPlus Files 3DPlus文件有问题

In 3DPlus, every data file you created was owned by lgc and had a user ID of 200 and a group ID of 10.
Now, in SeisWorks/3D you have the option of establishing different user and group IDs.

If you are using different accounts besides lgc with different user and group IDs, you may run into
permission problems when you try to access old 3DPlus files. Simply change the permissions to grant open
access. For instructions on checking and changing permissions, see “Changing the Permission Mode of a
File” below.

Changing the Permission Mode of a File 更改文件权限模式

Permissions are assigned to both directories and files. The permission mode of a directory determines
whether users can read or write to the files in the directory. The permission mode of a file determines
whether users can read, write to, or execute the process specified by that particular file.

Users belong to three different categories: (1) the owner of the file, (2) the group to which the owner
belongs, and (3) all others. If you are denied permission to read or write to the files in a directory, or to
read, write to, or execute a particular file, chances are that you must change a permission mode. You can
change the permissions assigned to a

particular directory or file only if you are the owner of the directory or file, or you have superuser status.
Examining the Currently Assigned Permissions 检查当前指定的权限

To view the current permission mode of a file, cd to the directory containing the file and use the ls -
lag command to display information about the files in the directory. You will receive a display that
resembles the following screen.

实例:

blade{owuser:/home/blade/owuser}% ll

-rw-rw-r-- 1 owuser staff 7270850 Apr 3 15:50 hjq-master

lrwxrwxrwx 1 owuser staff 17 Apr 18 2006 hjq_cgm -> /home1/pb/hjq_c

drwx------ 2 owuser staff 512 Mar 3 2006 Mail/

-rw-r--r-- 1 owuser staff 138985 Apr 10 09:14 zmapplus.zlog

blade{owuser:/home/blade/owuser}%

Column 1(红) indicates whether the file is a directory, a file, or a link. A dash (–) indicates that the file is a
file. A “d” indicates that the file is a directory. An “l” indicates a link to a file or directory.

Columns 2 through 4(蓝) indicate the permissions for the user. Columns 5 through 7(绿) indicate the
permissions for the group to which the user belongs. And Columns 8 through 10(粉) indicate the
permissions for all other users.

To determine what permissions are currently assigned to the different user types, use the following table.

Permission code Description

r Allows prospective user to read the file.

w Allows prospective user to write to or modify the file.

x Allows the prospective user to execute the file. This option only applies to files
that are executables.

- Denies access of the particular type indicated by the respective column. That is, if
a dash (–) appears in a read column, the read permission is denied. If it appears in a
write column, the write permission is denied. If it appears in an execute column,
the execute permission is denied.

Customizing Permissions for Different User Types为不同用户定制权限


You can change the permissions assigned to particular user types with the following chmod command.

chmod <user type><operator><permission code><filename>...

The variables that can be used in this command are described in the table below.

Variable Description

<user type> Specifies which of the user types is to be affected by the permission
change. Valid codes for user types: u is the owner, g is the group, and
o is all others.

<operator> Specifies whether permissions are to be granted or denied. Valid


options: + indicates that the permissions are to be granted; -indicates
that the permissions are to be denied.

<permission code> Specifies the type of permission to be granted or denied. Valid codes:
r applies to read privileges, w applies to write privileges, x applies to
execute privileges.

<filename> Names the directory or file(s) to which the permissions are to be


applied. If <filename> is the name of a directory, the chmod
operation determines which users can read or write to the files in the
directory. If <filename> is the name of a file, the chmod operation
determines which users can read, write to, or execute the file.

Granting Read, Write, and Execute Permissions to All User Types为所有用户类型授予


读、写和执行权限

If you are the owner of a directory or file, you can grant read, write, and execute permissions to all user
types by using the following command

chmod 777 filename...

This method uses the octal number 777 to assign permissions. The octal number method of assigning
permissions is described in the online “man pages” for chmod (which you can view by typing man chmod
at a system prompt).

To change the permissions within a particular directory and apply the changes to all subdirectories and their
files, you can use the following “recursive” chmod command.

chmod -R 777 filename...

Problems with Horizons in SeisWorks SeisWorks中的层位问题


The following techniques can help you resolve problems with SeisWorks horizons or horizon files.

Cannot Make Another Horizon 无法生成新层位

If you receive an error message notifying you that you cannot make another horizon:

• Check for duplicate horizon data (.hzd) files. If you find duplicates, delete them.(删除重复的层位文
件)

• Check that the horizon catalog (.hrz_cat) exists. If it does not, create one with the crtcat utility.

• Check whether you have exceeded the limitation of 4096 horizons currently in the working project. If you
have, you must use an existing horizon plane, since the 4096 horizon limit is arbitrary and irreversible.
Eliminating existing horizons will not enable you to add additional horizons. Instead, find a horizon that
you no longer need. Display the horizon in the Map View and use the Areal Delete option to delete the
interpreted data that is currently correlated with the horizon. Use the Attributes option in the

Seismic View to rename the horizon. Reinterpret the horizon.

Horizon List Incorrect or Missing 层位列表错误或丢失

SeisWorks/2D

First, check to see if the horizons are still in the master project but simply not appearing in your working
project. Select Horizons à Global Manager à Global Add to Project. If the horizons you need are
listed, select them for your working project. If the horizons of interest are not listed in the Global Add to
Project dialog box, check to see if both horizon header(层位头) (.hzh) and horizon data(层位数据) (.hzd)
files are contained in the master project. If the data files are present but the header file is missing, simply
“creating” the horizon once again within SeisWorks may solve the problem. But if the data files are
missing, you will have to reinterpret the horizon or load it from a backup tape.

(例如,在master目录下:

T8h______________________________________.hzh_glb

t8h_______________________________________.hzh_glb

HE-2000-119H______________T8h____________.hzd_glb

HE-2000-119H______________t8h_____________.hzd_glb )

Check the dir.dat file to make sure that all the master project filesystems are in directories with the “global”
designation. If these filesystems have somehow been moved to directories that are not global, you cannot
access the files you need to display and work with horizons.

(例子:
blade{owuser:/home/OpenWorks/conf}% more dir.dat

/home/OWPROJ sys global

#/home1/OWPROJ global

/home4/OWPROJ global

blade{owuser:/home/OpenWorks/conf}% )

Move the master project filesystems to a global directory.

If dir.dat is not properly specified or your global area is not properly defined for the project, the horizons
will be marked as deleted in the horizon catalog (.hrz_cat). Correct the problem with dir.dat or the global
area, and then simply use the Global Manager to restore your horizons to the working project.

SeisWorks/3D

If you find some of your horizons missing, try rebuilding the horizon index file (.hrz) using the HrzUtil
utility. See the HrzUtil chapter of the Seismic Project Utilities manual for detailed procedures.

1. Exit SeisWorks/3D.

2. In an xterm window, enter the following command:

HrzUtil

3. When the HrzUtil dialog window appears, select the 3D radio button and click on List. From the popup
Project Selection dialog box, highlight your project and click OK.

A list of horizon names, ordered by horizon number, appears in the Horizon List selection box.

4. Select Rebuild Index.

5. When the index has been rebuilt, the horizon list is refreshed.

Examine it to confirm that the missing horizon has been recovered.

Select Alphabetic if you prefer to view the list in alphabetical order.

6. Select Quit and restart SeisWorks/3D.

Displaying Horizons in Map View 在Map View中显示层位

You can have a total of five active horizons for all currently open Map or Perspective Views. Your horizon
selections apply to all currently open Map Views, Horizon Image Maps, and Perspective Views.

For example, suppose you select bluemax as the first display horizon in the Contents dialog box for Map
View 1. Then you select greenmin as the first display horizon in the Contents dialog box for Map View 2.

Your last selection will be applied to all the currently open views. If you redraw Map View 1, greenmin—
not bluemax—will be displayed there.

To avoid this problem, use the second horizon display field when you choose the horizon for your second
Map View, the third horizon display field when you choose the horizon for your third Map View, and so on.

Problems with Faults in SeisWorks SeisWorks中断层问题


Below are some techniques to help resolve problems in working with SeisWorks faults.

Correlating Faults 关联断层

The following tips address problems with correlating faults in SeisWorks.

Fault Plane Not Listed 未列出断层面

Only faults that have been selected for display are listed in the Correlate dialog box. If a fault you want is
not there, close the dialog box. From the Map View or Seismic View menu bar, select Faults à
Select, and choose that fault for display. Then reopen the Correlate dialog box.

Cannot Assign Fault Heave 无法指定断层高点

When you are correlating with fault heaves in Map View, you must click on the fault segment within that
heave. The fault heave, which represents the gap in the horizon, may be wider than the fault segment in
areal view. So when you click within the heave you may not actually be on the segment, as illustrated
below.(图略)

Fault as interpreted in Seismic View. Note that gap in horizon is wider than the area covered by the fault segment as drawn.

Fault as shown in Map View. Note that the width of the heave exceeds the width (areal extent) of the fault segment. When you
correlate, you must click on the segment.

Relation of Fault Heave to Fault Segment

If you click within the heave and get a beep, move closer to one side of the heave symbol and click again.

Identifying Faults in Map View 在Map View中识别断层

In a Map View, the Identify Faults option works only for faults that have been triangulated. Any change to
a fault segment or fault plane makes the existing triangulation of that plane obsolete.

If the triangulation is out of date when you try to identify faults, a message prompts you to retriangulate or
exit the identify faults mode.

To retriangulate, click on OK. The program retriangulates the fault plane and continues the search to
identify faults.
This message will also appear if only one fault segment has been assigned to the fault plane. To avoid
receiving the message, digitize another fault segment and apply it to the fault plane.

Suspect-Looking Fault Contours 不真实的断层等直线


Triangulation connects the digitized points of all fault segments assigned to a particular plane. If you have not picked those segments consistently, the
triangulation will be irregular and inaccurate. As a consequence, the fault contours will also be irregular and inaccurate.

To avoid this irregularity, keep all segments for a given plane approximately the same length. (That is, digitize the full extent of the fault segment on
every section you interpret.)

Problems Exporting Data from SeisWorks从SeisWorks输出数据


问题
With the Export to Z-MAP Plus option, exporting a large horizon can take considerable time. If an export
job fails or appears to fail, use the following strategy to resend your export job.

1. Check to see if the process is still running by typing

ps -ef|grep SNAP (IBM, Solaris, and SGI)

2. If the export job is not running, perform the following steps:

• Check to see if your filesystem has run out of space.

Export to Z-MAP Plus creates temporary sorting files (tmp.*) in the project sys directory during an export
job. Your horizon may exceed the available capacity.

• Remove the tmp.* files.

• Decimate the horizon and convert it to a dataset using the

Convert Horizon to Map Points option.

• Export the horizon dataset using the Computed Contours and Fault Polygons option on the Export to Z-
MAP Plus dialog box.

Checking the Continuity of SeisWorks/2D Data Files 检查


SeisWorks/2D数据文件连续性

For a quick check of your 2D master project, use the contest2d utility.

This utility tests for continuity of 2D line data, horizon data, and seismic data.

It checks to see if

• each line has a a valid trace range(每条测线拥有有效道范围). That is, it checks that the minimum trace
is not 0, that the minimum trace number is smaller than the maximum trace number, and that the minimum
and maximum trace numbers are not the same.

• each line has the required navigational files(每条测线拥有所需的导航文件). These are the .st_glb file
that defines the shotpoint-to-trace relationship(炮道关系) and the .xy_glb file that contains shotpoint real-
world locations(炮点实际位置).

• every seismic file (.2v2)(地震文件) has read/write permissions.

• every horizon data file (.hzd_glb)(层位数据文件) has read/write permissions.

• every horizon data file (.hrd_glb) has the same trace range as defined in corresponding the line header file
(.lh_glb)(测线头文件).

As contest2d runs, it outputs to the screen a list of each line examined and the trace range of that line. The
utility also generates two files:

• a log file that lists every file examined, its file size, and its full path

• an error file that lists any discrepancies encountered during the check.

The log file is useful in compiling a database of the lines in your master project and in keeping track of the
size of your master project.

The error file may help you identify and resolve problems encountered in displaying seismic or horizon
data.

See the contest2d chapter of the Seismic Project Utilities manual for detailed procedures.

To run contest2d, do as follows:

1. Set the environment variable LGC_MASTER to the master project you wish to examine by entering the
appropriate commands at the prompt:

• In a C shell: setenv LCG_MASTER <master_project>

• In a Bourne shell: LGC_MASTER=<master_project>

export LGC_MASTER

2. At the prompt, type

contest2d <openworks project name> <contest log file > < error file >

Use whatever names you like for the log and error files. If you want to write them to a directory other than
the current directory, include the full path.

3. Examine the output on the screen, which will show you each line checked and its trace range.
4. Examine the log file and error file.

Data Loading Problems 数据加载问题


This section provides solutions to common problems that can occur after loading data into an OpenWorks
project.

General Data Loading Problems 普通数据加载问题

TVD Total Depth for Deviated Wells is Too Shallow 偏移井TVD总深度太浅

Try one or more of the following:

• Load a position log with TVD values.

• Load a position log with more stations (depths) recorded along the well path.

• Load a directional survey (it will calculate TVD values using the radius of curvature method as the
default).

Message: Records dropped due to errors

This message does not necessarily indicate a problem. When dealing with multiple occurrences of a field,
null occurrences can exist after the last data value. The loader cannot write a record with a required field set
to null, so it drops that record. Check the error log file to verify.

SeisWorks Data 数据显示问题

Wells Appear in 2D Display but not in Map View (SeisWorks)

Try exiting and restarting SeisWorks. If that does not correct the problem, try one or more of the following:

• Make sure that the active OpenWorks project matches the SeisWorks definition.

• Check x, y coordinates; modify as necessary using SeisWorks World Coordinates utility.

• Reload well x,y coordinates using a different cartographic reference system.

Well Symbols Appear in 2D Display but not in Map View

The Well Status is not recognized by SeisWorks. Edit the status using the OpenWorks Well Data Manager.

Wells Appear in Map View but not in Seismic View

You are missing one or both of the following: a Time/Depth Table, a Position Log.

Wells Appear in Seismic View, but not Curves or Tops


You could be missing a Time/Depth Table or using a bad table. You might also need to edit the synthetic to
fix the time-depth relationship.

Geologist’s Tops not Appearing Correctly in Seismic View

Do one or more of the following:

• Check the SeisWorks project datum settings.

• Make sure that KBs are loaded for the wells.

• Edit the synthetic to fix the time-depth relationship.

Synthetics do not appear in Seismic View

Check to see that Synthetics are toggled on in Seismic View:

1) Contents-Log/Synthetic Track; 2) Wells-Parameters-Display Curve;

3) Wells-Select-Synthetics. Also make sure that the synthetic exists over the time range displayed. Often,
synthetics have a time reference other

than zero depth and require bulk shifting.

Troubleshooting Seismic Display Problems


排查地震显示问题
This chapter contains solutions to problems you might encounter with displays of seismic data. These
problems could include gaps in seismic lines, color allocation problems, or bad displays of computed
contours or wells.

What’s in this Chapter 本章内容


This chapter contains the following sections:

• Problems with Seismic Display on page 58. Contains solutions for common problems encountered while
displaying seismic data.

• Problems with Colors on page 59. Describes steps you can take to solve problems related to the allocation
of color for seismic displays.

• Other Display Problems in SeisWorks on page 60. Discusses solutions for problems with displaying
computed contours or wells in map displays.

Problems with Seismic Display 地震显示问题


Some suggestions for managing problems with your seismic display are discussed below.

No Seismic 无地震剖面显示

SeisWorks/2D

Did you select a vertical seismic (.2v2) file? Check the Seismic Parameters dialog box, and select a .2v2
file. (是否选上了 .2v2 文件)

SeisWorks/3D

If no seismic is displayed at all, you may need to select a vertical, bricked, or compressed seismic (.3dv, .
bri, .cmp) file. Check the Seismic Parameters dialog box, and select a seismic file. (是否选上了 .3dv, .
bri, .cmp 文件)

Snow or Stripes on Display 剖面显示上出现雪花或条纹

If you are trying to display a 16-bit or 32-bit file, and you have not scaled the file to 8 bits, you will receive
“snow” or stripes on the display.(显示16或32位文件时,没有按8位模式来显示)

See “Scaling and Clipping Seismic Data” in the chapter “Preparing Seismic Views” of the SeisWorks/3D
Data Display manual for information on how to scale seismic data.

Gaps or White Spots on Seismic Line 测线上出现缺口或白点

Gaps or white spots may indicate problems in the way the data was acquired or processed. Another
possibility is that the data was not loaded correctly. If the problem is extensive, you may have to reload the
data. Check with your system administrator or data loader. Also, refer to the 2D Project Management, 3D
Project Management, 2D Batch Control Monitor, and 3D Batch Control Monitor manuals.

Problems with Colors 颜色问题


Often what appears to be a data problem is simply a problem in displaying that data.

Can’t Get Dual Color Bars 无法调用双色棒

When you start SeisWorks, you are asked to indicate whether you want single or dual color bars on each
screen. If you get a message saying dual color bars cannot be allocated, you probably do not have enough
colors available. Remember other applications are also tapping the color resources.

In such a case, you can either run SeisWorks with a single color bar, or you can kill some other applications
and then try to start SeisWorks in the dual color bar mode.

No Data Displayed 无数据显示

If you display a horizon or mapping file in the Map View, and nothing appears, display the Color Control
dialog box of the screen where the problem is occurring.

The problem may be that the Mode option is set to Manual and the data is outside the range of the color
bar. If this is the case, set the Mode option to Local and redisplay the horizon or mapping file in the Map
View in question.

No Color Maps 颜色文件丢失

Color map files (.clm files) are automatically copied to the project sys directory when you create a working
project. Select FileàOpen in the Color Control dialog box to see whether you have color map files. If
none are listed in the resulting dialog box, someone has deleted them or moved them out of the system. To
restore them, copy the .clm files from $SEISUTILSHOME/dat to your project sys directory.

(例如单位上的机器:

.clm 默认存放位置在:/home/OpenWorks/SeisUtils/dat

project sys 位置在:/home/OWPROJ/OW_SYS_DATA/color_files

2D工区的存放位置(包括自定义的颜色文件):/home/OWPROJ/hjq072d)

Partial Data Displayed 只显示部分数据

If the color bar is not picking up the minimum and maximum values properly, recompute extremes for the
horizon. In a Map View or Horizon Image Map, select Horizons à Computations àCompute
Extremes. Then set the color bar to the Local mode and redraw the Map View.

Be sure that in specifying the areal extent for the Compute Extremes computation, you select Entire
Survey. If Compute Extremes is run for a limited map area, the color bar will reset to the minimum and
maximum values computed for that small area only.

Other Display Problems in SeisWorks SeisWorks中其它显示问题

This section contains troubleshooting techniques for computed contours and well data in map displays.

Troubleshooting Computed Contours 等值线问题

The quality and accuracy of computed contours depends on the contouring parameters you have set. If you are
dissatisfied with their appearance, adjust the contour parameters and rerun the contouring job.

Troubleshooting Computed Contours 排查等值线计算问题

Problem问题 Possible Solutions可能的解决办法


Computed contours do not appear • Be sure contours are turned on in the Map View
in Map View. Contents dialog box.

图中未显示出等值线 • Check the Color Control contour settings to


determine whether the contour range falls within the
map point range.

Use the View option (Mapping FilesàView) to do


the following:

• Make sure that you have opened the appropriate


mapping file.

• Check the mapping file to see that the grid exists.

•Check the grid parameters to determine whether the


search radius is too small.
Too few or too many contours. 等 • Reset the Start Time, End Time, and/or Interval settings
值线太密或太稀 in the Color Control window and then redraw the Map
View.
Gridded contours are too noisy or • Increase grid size and search radius.
jagged. 网格不合理

Gridded contours are • Increase search radius.


discontinuous (broken) 网格不连

Troubleshooting Computed Contours 等值线计算问题

Problem 问题 Possible Solutions 可能的解决办法

Contours appear inside the fault • Use the Edit Polygons option to delete the points inside
polygons. 等值线在断层多边形 the polygons.
内部
• If the contours are gridded, check the parameters you
set in Map It. If you selected Do not use polygons,
recalculate the gridding surface using the Use existing
faults option.
• Check the following in the Computed Contours
No annotation appears on Parameters dialog box (ContoursàParameters
Computed):
the contours. 等值线上无标注
• Computed Contour Annotation is turned on.

• The text size is big enough.

• The text size is not too big.

• The annotation interval coincides with the times


contours are drawn.
Troubleshooting Your Well Display井显示问题

Frequently problems encountered while displaying wells are related to your parameter settings or your time-
depth tables. Use the following table to check that you have set the well display parameters properly for the type
of data you want to view.

Problems with Well Display 井显示问题

Problem问题 Possible Solutions可能的解决办法

Reset the Criterion Distance option in the Seismic


Well Parameters dialog box to a larger value.
Check that the missing wells have been selected for
display and that Generate well list for each new
section is properly set.
Check that the missing wells have time-depth
Wells do not appear on the Seismic tables assigned to them.
View. Seismic View上无井显示 Check that a measured total depth has been
specified in the Well Data Manager utility.
Check that position logs exist for the missing wells.
Use the Well Data Export utility and attempt to
export the well’s position log. If the exported file
is empty, the well has no position log.

Check that the well’s x,y coordinates are


specified in the same units of measurement as the
project. Use the Well Data Manager.

Problems with Well Display 井显示问题

Problem问题 Possible Solutions可能的解决办法

Too many wells are posted. • Reset the Criterion Distance option in the
Seismic Well Parameters dialog box to a smaller
value.

• Remove the wells that you do not want to be


displayed from the active wells list.

• Set the Well Display Mode option in the Seismic


Well Parameters dialog box to Part. Use the Show
Location option to identify the portions of wells
that now appear in the Seismic View.
The curves that you expect to be Check that the Display Curve option in the
displayed do not appear. Seismic Well Parameters dialog box has been set to
Amplitude Log.

Check that the curve has been selected for display.

Reset the Track Width option in the Seismic Well


Parameters dialog box to a higher value.
Curves are overposted. Change the placement of the curves so that they
are evenly distributed on the left and right sides of
the well borehole.
Tops are not displayed. Check that tops have been turned on in the
Contents dialog box and that the tops to be
displayed have been selected using the Tops
option under Wells on the menu bar.

Troubleshooting System Problems 排查系统问题


This chapter explains how to troubleshoot errors in the OpenWorks system and provides some recommended
courses of action.

What’s in this Chapter


This chapter contains the following sections:

• Checking OpenWorks System Processes on page 64(检查OpenWorks系统进程). This section describes


the system processes that must be running for your OpenWorks system to function properly and contains
procedures for starting and stopping these processes.

• Troubleshooting System Problems on page 68(排查系统问题). This section contains solutions to a few
common Landmark system problems.

Checking OpenWorks System Processes 检查OpenWorks系统进程


OpenWorks uses the following system processes which may be running on the workstation. All of these
processes are required.

oracle Oracle Server. The oracle server processes must be running


at all times on the workstation. (全程运行)
pd Pointing Dispatcher. Handles data communication between
OpenWorks applications. Is started dynamically for each
OpenWorks session.(调度程序. 在OpenWorks应用程序间
管理数据通信)
NetD Network Services for OpenWorks applications.(针对
OpenWorks应用程序的网络服务)
viewer The Help system software. Requires X to be running. (帮助
系统软件。要求运行X)
lmgrd License Manager Daemon. This only runs on the machine
which is the LAM manager. Typically started at boot time.
(许可管理器守护程序)

licsrv License Server Process. This only runs on the machine


which is the LAM manager. Typically started at boot time.
(许可服务器进程)

To check if OpenWorks processes are running, use the ps command with the -l option. For example:

ps -efl
(IBM, Solaris, or SGI)

To check on Oracle processes, use the su command to switch to the Oracle user account, then use the ps
command. For example:

su - oracle

ps -l
(IBM, Solaris, or SGI)

(例子:

blade{owuser:/home/blade/owuser}% su - oracle

Sun Microsystems Inc. SunOS 5.8 Generic Patch October 2001

blade% ps -l

F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME ***

8 R 105 3030 2943 0 41 20 ? 176 pts/12 0:00 csh

blade% pwd

/home/oracle

blade% )

To check on a specific process, use the grep option with the name of the process or a fragment of the process
name. For example:

ps -efl | grep ora


(IBM, Solaris, or SGI)

(单位机器上的例子:

blade% ps -efl | grep ora


8 S oracle 537 1 0 71 20 ? 1156 ? 07:23:19 ? 0:00 /home/oracle/OraHome1/bin/tnslsnr L

8 S oracle 566 1 0 40 20 ? 23423 ? 07:23:23 ? 0:00 ora_pmon_owblade

8 S oracle 568 1 0 40 20 ? 23422 ? 07:23:23 ? 0:00 ora_dbw0_owblade

8 S oracle 570 1 0 40 20 ? 23359 ? 07:23:23 ? 0:00 ora_lgwr_owblade

8 S oracle 572 1 0 41 20 ? 23350 ? 07:23:23 ? 0:00 ora_ckpt_owblade

8 S oracle 574 1 0 41 20 ? 23380 ? 07:23:23 ? 0:00 ora_smon_owblade

8 S oracle 576 1 0 41 20 ? 23348 ? 07:23:23 ? 0:00 ora_reco_owblade

8 S oracle 578 1 0 40 20 ? 23339 ? 07:23:23 ? 0:00 ora_snp0_owblade

8 S oracle 580 1 0 41 20 ? 23369 ? 07:23:23 ? 0:00 ora_snp1_owblade

8 S oracle 2182 1 0 41 20 ? 23672 ? 07:50:46 ? 0:01 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2180 1 0 41 20 ? 23360 ? 07:50:46 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2246 1 0 41 20 ? 23602 ? 07:51:15 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2244 1 0 41 20 ? 23348 ? 07:51:15 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2415 1 0 41 20 ? 23348 ? 07:56:11 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2421 1 0 41 20 ? 23367 ? 07:56:14 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2488 1 0 45 20 ? 23348 ? 07:58:41 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2490 1 0 55 20 ? 23349 ? 07:58:42 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2585 1 0 41 20 ? 23348 ? 08:17:44 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 2587 1 0 41 20 ? 23601 ? 08:17:44 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 3030 2943 0 41 20 ? 176 ? 08:48:10 pts/12 0:00 -csh

8 S oracle 3133 1 0 43 20 ? 23348 ? 08:48:37 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 3135 1 0 41 20 ? 23349 ? 08:48:38 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 S oracle 3312 1 0 41 20 ? 23348 ? 08:59:24 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n

8 R oracle 3370 3030 0 41 20 ? 120 09:16:39 pts/12 0:00 grep ora

8 S oracle 3314 1 0 41 20 ? 23372 ? 08:59:24 ? 0:00 oracleowblade (DESCRIPTION=(LOCAL=n


blade% )

Starting/Stopping System Processes 运行/中止系统进程

Occasionally, you may want to start or stop various system processes.

The following utilities are provided to help you.

Starting/Stopping the Pointing Dispatcher 运行/中止PD

The Pointing Dispatcher (pd) is started automatically through the Pointing Dispatcher Executor (PdEx)
daemon when Dynamic PD is turned on, which is the OpenWorks default setting. The Pointing Dispatcher
should be stopped only if you have a “dual environment”running multiple releases of OpenWorks
simultaneously. If you need to

run the stop_pd script, refer to “Configuring the Pointing Dispatcher”in the manual OpenWorks System
Administration: Managing the
System.

Starting Oracle 起动 Oracle

If you stop Oracle and do not reboot, you must restart Oracle before continuing to use OpenWorks. To restart
Oracle:

1. Use the su command to switch to the Oracle user account. For example:

su - oracle

2. Restart the database server by entering the following command:

dbstart

3. Start the Listener process by entering the following command:

lsnrctl start

Stopping Oracle 中止Oracle

Oracle keeps the most recent database changes in a cache. The contents of this cache may be lost if you reboot
the system without stopping Oracle. Consequently, you should always stop Oracle first before rebooting the
system. To stop Oracle:

1. Use the su command to switch to the Oracle user account. For example:

su - oracle

2. Shutdown the database server by entering the following command:

dbshut
3. Stop the Listener process by entering the following command:

lsnrctl stop

Stopping and Restarting Oracle from Server Manager

Both dbshut and lsnrctl work only if the letter Y appears at the end of the

configuration setting in the file /etc/oratab. If these processes do not work, you can
control the database server from inside the Server Manager utility using the
appropriate commands. See the Oracle Server Manager User’s Guide for more
details.

Starting the LAM Manager 起动LAM 管理器

To start the LAM manger, run the following script:

su - root

setenv OWHOME OpenWorks_home_directory

$OWHOME/lam/bin/startlmgrd

This should only be run on the machine which is the LAM manager.

Stopping the LAM Manager 中止LAM管理器

To stop the LAM manager, run the following script:

su - root

setenv OWHOME
OpenWorks_home_directory

$OWHOME/lam/bin/stoplmgrd

This should only be run on the machine which is the LAM manager.

Processes Initiated on Startup 起动初始化进程

When you start OpenWorks in a non-CDE environment using the startserver command, the following
processes are started: X, mwm, xsrm, viewer (FrameViewer), and launcher. When you start OpenWorks using
the startow command in an X window, the following

process is started: launcher.

The foregoing processes have the following dependencies:


pd Requires NetD to be running; NetD spawns产生
PdEx, which starts pd.
launcher Requires X to be running.

xsrm Requires X to be running.

viewer Requires X to be running.

mwm Requires X to be running, in a non-CDE


environment. 在非CDE环境上

Modifications to System Files 对系统文件的修改

Some system files were modified during the OpenWorks installation. If there are problems, check to make sure
that the correct modifications are present in the system files. The script $OWHOME/bin/rc.ow is the shell script
that starts most required processes when the workstation boots. The script rc.ow is called at boot time by the
system file /etc/inittab (IBM) or /etc/rc2.d (SGI and Solaris). The rc.ow start-up call

can be added to the system file when OpenWorks is initially installed.

Appropriate messages are displayed at the end of the boot stream if these files are set up properly. For example:

Starting Landmark License Server (LAM)...

NOTE: $OWHOME is an environment variable defined in rc.ow. This variable is also defined in .lgcprofile or .
lgclogin, where it should be the same as in rc.ow.

Troubleshooting System Problems 排查系统错误


The following pages contain guidelines for troubleshooting various system problems you may encounter while
running OpenWorks.

Troubleshooting X 排查X

Symptoms: No gray screen. X errors issued to console shell.

Solaris Workstations Solaris 工作站

1. Echo the value of the environmental variable $OPENWINHOME. This specifies where X has been loaded
(echo $OPENWINHOME).

2. Verify that files exist in that directory. Compare the file listing with a machine that is operating correctly.

Solution: If there are missing files, the best solution is to reload the OpenWindows software from the
OpenWorks release media. Refer to the “Installing OpenWindows” booklet.

Sun, Silicon Graphics, and IBM Workstations Sun等工作站

Check to see that all entries in the $HOME/.xinitrc file are valid.
Problems Starting SeisWorks SeisWorks起动问题

Below are solutions to some problems you might encounter in starting SeisWorks.
清理SeisWorks地震工区系统目录下无用文件的方法

/tmp Directory Full /tmp目录满了

Several files are created in the /tmp directory as you run SeisWorks. If these files become corrupted and are not
deleted when you reboot, you may have problems when you next attempt to open SeisWorks. To fix this
problem, look for the files in /tmp, delete them, exit SeisWorks, and then start SeisWorks again. You can safely
delete any of the following files. 当运行SeisWorks时,在/tmp目录下会生成一些文件。如果这些文件损坏
了并在重起机器时没被删掉,那么在打开SeisWorks可能会出现问题。为解决该问题,找到/tmp目录
里的这些文件并将其删除,退出SeisWorks,然后再重起SeisWorks。

.dirdat

.fs.lst

.plist.tmp

swild.ctr

upd.prj

*zycor.log (Export to Z-MAP Plus log)

“sys” Directory Full “sys”目录满了

During an interpretative session, SeisWorks creates certain temporary files and writes them to the project sys
directory. Under normal circumstances, these files are cleared when you exit SeisWorks. If they are not deleted,
you may find your sys directory becoming full. To remedy the situation, you can remove any of the following
file types. 在解释过程中,SeisWorks会生成一定类型的临时文件并将其写进地震工区系统目录中。在
正常情况下,当退出SeisWorks时这些文件会自动清除。如果没有清除,你会发现系统目录变得越来
越大(满了)。

hzbf.w# horizon buffers 层位缓冲文件

v<host_ID> map buffer files map缓冲文件

<host_ID>.w3s temporary session files 临时session文件

tmp.* Z-MAP Plus sorting files Z_MAP 排序文件

我在单位机器上用以下设置(在用户下以root身份用vi建个文本文件dk,加进如下内容保存,然后chmod
555 dk。执行dk即可清除/tmp目录和地震工区目录下无用的文件。用同样的方法,也可把个人的Z-Map作图
目录加进来,删除诸如 .lck, .LCK, tmp.*, core 文件。如果能编个角本程序当每次退出SeisWorks时能自动运行
并清理垃圾文件,将是一个高效的方法。):

rm /tmp/*.dirdat

rm /tmp/*.fs.lst

rm /tmp/*.plist.tmp

rm /tmp/swild.ctr

rm /tmp/upd.prj

rm /tmp/*zycor.log

rm /home/OWPROJ/hjq2d/v834d2bc2*

rm /home/OWPROJ/hjq2d/v834d2bc2*.w3s

rm /home/OWPROJ/hjq2d/hzbf*.w*

rm /home/OWPROJ/hjq2d/tmp.*

rm /home/OWPROJ/hjq2d/core

rm /home4/OWPROJ/hjq072d/v834d2bc2*

rm /home4/OWPROJ/hjq072d/v834d2bc2*.w3s

rm /home4/OWPROJ/hjq072d/hzbf*.w*

rm /home4/OWPROJ/hjq072d/tmp.*

rm /home4/OWPROJ/hjq072d/core

Troubleshooting Oracle Problems 排查Oracle问题


Introduction 引言
This chapter contains information to help you locate and resolve problems with the Oracle database. Use this
chapter to verify that Oracle has been installed correctly and has been configured to interact with the
OpenWorks software.

If you install Oracle using the Landmark Oracle install procedure, the procedure automatically configures
Oracle to run with OpenWorks. However, if your site already has Oracle installed, or you are going to install
Oracle directly from Oracle-supplied media, you must verify that your installation meets Landmark
specifications.

What’s in this Chapter 本章内容


This chapter contains the following sections:

• Before Using These Procedures on page 72. This section describes important information you should have
before troubleshooting Oracle.

• Documenting and Reading Error Messages on page 73. This section discusses the importance of recording
Oracle messages for use in troubleshooting.

• OpenWorks Requirements for Oracle on page 73. This section describes the settings and parameter values
required by OpenWorks for Oracle installations.

• The Environment Status Tool and Oracle on page 77. This section discusses the error messages that could
appear when using the Environment Status Tool to examine Oracle status.

• Using Server Manager to Query the Oracle Database on page 80. This section discusses the error messages
that could appear when using the Environment Status Tool to examine Oracle status.

• Using Server Manager to Query the Oracle Database on page 80. This section contains the procedure for
using Server Manager, a tool provided by Oracle for querying the database.

• Oracle in a Networked Environment on page 82. This section discusses important factors to consider when
troubleshooting Oracle in a networked environment.

• Troubleshooting Oracle Problems on page 85. This sections describes steps you can take to troubleshoot
common problems with Oracle.

Before Using These Procedures 排错前需知


The methods described in this chapter require database and system administration knowledge and skills.

The instructions in this chapter assume that OpenWorks has been installed successfully. Before reading the
chapter, you should be familiar with the following environmental variables.

• OWHOME — the location where the OpenWorks software is installed.

• ORACLE_HOME — the location where the Oracle software is installed.

• ORACLE_SID — the Server ID of the Oracle database instance.

Refer to the OpenWorks System Administration manual set for a description of these terms as well as for a
complete description of the OpenWorks environment.
In addition, you should be familiar with the OW_ADMINISTRATOR role that provides additional security for
OpenWorks projects. For a complete description of this role and the OpenWorks 2003 security model, see the
OpenWorks Project Management manual.

Documenting and Reading Error Messages 阅读并记录错误信息


When you encounter a problem, always document the error immediately. Whether it is an Oracle or application
error, you can help by maintaining a log of the errors and the subsequent solutions. Even if you need to get
technical support, you will have the problem well documented in a central log file for faxing or review.

Error messages prefixed by ORA come from the Oracle database directly or through the application system.
You have a simple facility available that can help identify the type of error. Enter the oerr command with the
message number, as follows. For example, if the

message number is 1234, you would enter:

oerr ORA 1234

The complete message is displayed below the command line for reference. This rarely provides a complete
answer or solution but can provide quick hints to solving the problem. If you need more help, you can check
your Oracle documentation for documented error messages, or call for help. If you call, be prepared to explain
the events leading up to the problem. You may also want to check your log file to find any previous errors that
may have led to your current problems.

OpenWorks Requirements for Oracle OpenWorks要求的Oracle及其补



OpenWorks 2003.3.0 requires Oracle 8.1.7.2, the 8.1.7.7 patch, and Net8 Configuration Assistant. The Oracle
instance must meet with Landmark specifications for OpenWorks to interact successfully.

If you install Oracle using the procedures described in the OpenWorks Installation Procedures document,
including using the pre-Oracle process, all the parameter settings described in this section should already have
been configured to interact with Landmark applications.

If Landmark media is not the source of your Oracle database, please follow the recommendations in this section.

Recommended Settings 推荐配置

The table below shows,recommended resource settings.

Resource Recommended

maxinstances 1

maxlogfiles 40
maxdatafiles 254

maxlogmembers 5

Recommended Datafile Sizes 推荐Datefile大小

Default datafile sizes provided by Oracle are inadequate for Landmark applications. The required datafile sizes
are as follows:

Tablespace Name Default Data File Namea Landmark Default


Size

System Tablespace sysowminnie.dbf 600Mb

First Redo Log Files logg1m1owminnie.dbf 8Mb 8Mb


logg1m2owminnie.dbf

Second Redo Log Files logg2m1owminnie.dbf 8Mb 8Mb


logg2m2owminnie.dbf

Third Redo Log Files logg3m1owminnie.dbf 8Mb 8Mb


logg3m2owminnie.dbf

Rollback Segment rbsowminnie.dbf 200Mb

Temporary Tablespace tempowminnie.dbf 100Mb

Users Tablespace usrowminnie.dbf 10Mb

Tools Tablespace toolowminnie.dbf 15Mb

a. Oracle appends the ORACLE_SID to the filename. In the example, the ORACLE_SID is owminnie, where minnie is the server
name.

For safety purposes Landmark strongly recommends mirroring Oracle Log files on separate disks. However, not
having mirrored Log files will not affect the functioning of the Landmark applications.

Items Affecting Size Requirements 影响Datefile大小的因素

The sizes shown above for system and rollback tablespaces may be insufficient if your site experiences any of the
following conditions:

System Tablespace

Each OpenWorks project will take from 3 to 5 Megabytes of space in the system tablespace.

Rollback Tablespace

You may find that you wish to increase the size of your rollback tablespace if you encounter problems in doing
the following:

• Creating large grids or pointsets in Z-MAP Plus or StratWorks.


• Working with very finely sampled log curve data in PetroWorks.

• Doing project backups while users are working in the very large projects.

• Doing backups for projects that have a very large number of wells (30,000 or more).

Database Block Size

Landmark recommends a default database block size of 8 Kilobytes.

Oracle sets a default block size of 2 Kilobytes. Set the parameter in the instance configuration file config.ora to
the following.

db_block_size = 8192

Oracle Parameters

Landmark applications have a few requirements about values of certain Oracle parameters. The following lines
must appear in the instance initialization file, commonly known as the init{ORACLE__SID}.ora file.

The db_files and max_enabled_roles parameters can have the given values or larger values.

compatible=8.1.7.1

os_authent_prefix=""

remote_os_authent=true

#NLS_DATE_******="YYYY-MM-DD HH24:MI:SS"

db_files=254

max_enabled_roles=48

shared_pool_size=9000000

db_block_buffers=3200

The following lines are recommended for performance reasons (the job_queue items are required for
OpenExplorer Advanced Project Management):

pre_page_sga = true

ccf_io_size = 131072

sort_area_size = 524288

#small_table_threshold = 20
open_cursors = 400

log_buffer = 1048576

max_dump_file_size=10240

dml_locks=500

job_queue_processes=2

job_queue_interval=60

Applying the Changes 使修改生效

After making the changes described above, you must shut down the database and restart it for the changes to
take effect. After the Oracle instance has been restarted, please ensure that the OpenWorks-supplied database
setup script is run by the Oracle DBA user. This script creates the tablespaces and Oracle users required by the
OpenWorks database,

Landmark specific views and other relevant steps.

The listener process must be to running before the setup script is run.

To start the listener process, as the Unix Oracle DBA user, type:

lsnrctl start

To run the database setup script, as Unix Oracle DBA user, type:

$OWHOME/install/owdbsetup

The Oracle role MONITORER must be defined in the Oracle instance.

Please have your site DBA run the Oracle supplied script utlmontr.sql as the Oracle user SYS.

The Environment Status Tool and Oracle 环境状态工具和Oracle


The OpenWorks Environment Status Tool allows you to examine various components of the OpenWorks
system, including the Oracle database. You may encounter error messages when you try to check the status of
Oracle. Following are examples of error messages that might pop-up if the Status Tool encounters problems.

Cannot open the database instance configuration file 无法打开数据库范例配置文件

This error is indicated by the following message.

错误信息显示窗口上显示为:

Cannot open the database instance configuration file


Please make sure that:

i) the Oracle instance is running for the ORACLE_SID defined

ii) the current user is added as a valid Oracle user to the database

iii) there is read permission for the instance configuration file.


OK

There are three possibilities, summarized in the message box and listed below.

• The Oracle instance is not running on the machine where the Environment Status Tool is being invoked.
This can typically happen in a client-server environment where the user is running the Environment Status Tool
on the Oracle client. The tool needs to run on the Oracle server.

• The Unix user running this tool is not a known user to the Oracle instance that is being queried. (The one
that is defined by the ORACLE_SID environment variable.)

OpenWorks provides a script that simplifies this process. To run the script login to the Oracle DBA account on
the machine where Oracle is running, and type:

$OWHOME/bin/orauser

Proceed to add the current Unix user as an externally identified Oracle user. Selecting all the defaults provided
by the script will do the necessary steps.

This process is typically done when a Unix user is created by the lgcuser script provided by OpenWorks to add a
new user or update an existing user.

• The instance configuration file, typically called $ORACLE_HOME/dbs/config.ora should be readable by all.

Some Oracle installations make this file read protected. The Environment Status Tool requires this file to be
readable by the Unix user running the tool.

Could not find Oracle Project Locations 找不到Oracle工区位置

If the Oracle database can not find your OpenWorks project directories, the following message appears.

错误信息显示窗口上显示为:

Could not find the Oracle Project Locations

Please ask your DBA to run the script $OWHOME/bin/dboradir and add project
locations.
OK

OpenWorks requires project locations to be known to the Oracle database. Project locations are directories
where the OpenWorks software will create database files associated with OpenWorks projects.

These locations are added when the script dboradir is run. If Oracle has been installed using the Landmark-
supplied owinstall format, this will automatically be done for you. Otherwise the Landmark-supplied script
owdbsetup needs to be run on the machine where the Oracle database is installed by the Unix Oracle DBA user.
If no project directory locations are specified, the default location of $ORACLE_HOME/dbs is

assumed.

Could not find the Oracle Database Log Files 找不到Oracle Database Log文件

If the Oracle role MONITORER has not been created, the following message appears.

错误信息显示窗口上显示为:

Could not find the Oracle Database Log Files

i) run the script

$ORACLE_HOME/rdbms/admin/utlmontr.sql from the SYS account.

ii) grant role MONITORER to the current user running the Status Tool.
OK

For the Environment Status Tool to correctly diagnose the Oracle configuration, the user running the tool
should have the Role MONITORER granted. Once again if the installation is a standard Landmark installation,
this is already done for you. If this is not a standard Landmark installation, the role MONITORER first needs to
be created for this particular Oracle installation.

Please have your Site DBA run the Oracle-supplied script utlmontr.sql as the Oracle user SYS. Once that is done
have the DBA assign the role to the Unix user running the Environment Status Tool. The utlmontr.sql script is
found in the $ORACLE_HOME/rdbms/admin directory.

Using Server Manager to Query the Oracle Database用服务器管理器查


询数据库
Server Manager is a tool provided by Oracle to query the database instance. This section describes how to use
Server Manager to query the database for certain information that you need to know.

OpenWorks requires three OpenWorks specific internal Oracle users to be created:

• owsys

• owsysp
• lmsys

The users’ owsys and owsysp have tablespaces owsys and owsysp associated with them. Use the following
procedures to check whether the users and tablespaces exist and are set up correctly.

Login as the Unix Oracle DBA user, then type the following:

svrmgrl

You will be put in the Server Manager environment. Type:

connect internal

You are now connected as a privileged Oracle user called internal.

Next, query the Oracle instance:

select username, default_tablespace,temporary_tablespace from dba_users;

You should see a listing similar to the following:

USERNAME DEFAULT_TABLESPACE TEMPORARY_TABLE

--------- ----------------------- -----------------


SYS ****** TEMP
****** TOOLS TEMP
OWSYS OWSYS TEMP
OWSYSP OSWYWP TEMP
LMSYS ****** TEMP
VISPI USERS TEMP
6 rows selected.
The users SYS and ****** are present in all Oracle instances. Notice that the users OWSYS and OWSYSP are
associated with the default tablespaces OWSYS and OWSYSP. The preceding query verifies that the relevant
users exist. If these users do not exist, you must run the OpenWorks owdbsetup
script as the Unix Oracle DBA user.

The next query verifies the existence of the tablespaces OWSYS and OWSYSP. Type this command:

select tablespace_name, status, file_name from dba_data_files order by


tablespace_name;

This query lists all the tablespaces known to the current Oracle instance, the status of these tablespaces, and the
data files associated with each tablespace. Your screen listing should be similar to the following output:

TABLESPACE STATUS ****_NAME

------------------------------------------------------

OWSYS AVAILABLE /pa/ow/oracle/dbs/owsys_tbs.dbf


OWSYSP AVAILABLE /pa/ow/oracle/dbs/owsysp_tbs.dbf

RBS AVAILABLE /pa/ow/oracle/dbs/rbsownova.dbf

****** AVAILABLE /pa/ow/oracle/dbs/sysownova.dbf

TEMP AVAILABLE /pa/ow/oracle/dbs/tempownova.dbf

TOOLS AVAILABLE /pa/ow/oracle/dbs/toolownova.dbf

USERS AVAILABLE /pa/ow/oracle/dbs/usrownova.dbf

7 rows selected.

The preceding list shows that the tablespaces OWSYS and OWSYSP exist and are available. It also verifies that
the data file associated with the tablespace physically exists on the system. If you find that the OWSYS and
OWSYSP tablespaces do not exist, run the OpenWorks script owdbsetup as the Unix Oracle DBA user.

If the tablespaces OWSYS and OWSYSP exist but their status is OFFLINE or INVALID, please have the site
DBA make them AVAILABLE or call your Landmark representative for assistance.

To exit from Server Manager, type:

exit;

More on Oracle users in OpenWorks Project Management manual

In addition to the basic roles described here, the OW_ADMINISTRATOR role that
provides additional security for OpenWorks projects. For a complete description of
this role and the OpenWorks 2003 security model, see the OpenWorks Project
Management manual, OpenWorks Security Model chapter.

以下是单位机器上的操作实例:

blade{owuser:/home/blade/owuser}% su – oracle 回车
Sun Microsystems Inc. SunOS 5.8 Generic Patch October 2001
blade% svrmgrl 回车

Oracle Server Manager Release 3.1.6.0.0 - Production

Copyright (c) 1997, 1999, Oracle Corporation. All Rights Reserved.

Oracle8i Enterprise Edition Release 8.1.6.1.0 - Production


With the Partitioning option
JServer Release 8.1.6.1.0 - Production

SVRMGR> connect internal 回车


Connected.
SVRMGR> select username, default_tablespace,temporary_tablespace from dba_users; 回车
USERNAME DEFAULT_TABLESPACE TEMPORARY_TABLESPACE

------------------------------ ------------------------------ ------------------------------

SYS ****** TEMP

****** TOOLS TEMP

OUTLN ****** ******

DBSNMP ****** ******

TRACESVR ****** ******

OWSYSP OWSYSP TEMP

OWSYS OWSYS TEMP

AURORA$ORB$UNAUTHENTICATED ****** ******

LMSYS OWSYS TEMP

OW_ADMIN_UTILS OW_ADMIN_UTILS TEMP

OWUSER USERS TEMP

OE_GRAPHS OE_GRAPHS OE_GRAPHS

HJQ HJQ HJQ

OWJH USERS TEMP

JH USERS TEMP

OWLYP USERS TEMP

OWLXL USERS TEMP

OWFJX USERS TEMP

WXR USERS TEMP

DB1 DB1 DB1

20 rows selected.

SVRMGR> select tablespace_name, status, file_name from dba_data_files order by tablespace_name; 回车

TABLESPACE_NAME STATUS ****_NAME

-------------------------- ------------ -------------------------------------------------------------------


DB1 AVAILABLE /home/oracle/OraHome1/owprojects/DB1022329.dbs

HJQ AVAILABLE /home/oracle/OraHome1/owprojects/HJQ08139.dbs

INDX AVAILABLE /home/oracle/OraHome1/oradata/owblade/indx01.dbf

OE_GRAPHS AVAILABLE /home/oracle/OraHome1/owprojects/OE_GRAPHS7900.dbs

OWSYS AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/owsys_owblade.dbf

OWSYSP AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/owsysp_owblade.dbf

OW_ADMIN_UTILS AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/ow_admin_utils_owblade.dbf

RBS AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/rbsowblade.dbf

****** AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/systemowblade.dbf

TEMP AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/tempowblade.dbf

TOOLS AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/toolsowblade.dbf

USERS AVAILABLE /home/oracle/OraHome1/database/datafiles/owblade/userowblade.dbf

12 rows selected.

SVRMGR> exit 回车

Server Manager complete.


blade% pwd 回车
/home/oracle
blade% exit 回车
blade% logout
blade{owuser:/home/blade/owuser}%
(注:以上的******号处为 S Y S T E M,因为本论坛禁用该字符。)

Oracle in a Networked Environment 网络环境中的Oracle


Oracle in a networked environment could imply different configurations.

• peer-to-peer configuration(对等网配置). This means that two or more Oracle instances, on separate or
same machines, each having stand-alone capabilities, are set up to communicate with each other.

• client-server configuration(客户机-服务器配置). This typically means a machine is

NFS mounting Oracle binaries from an NFS server, and accessing an instance on a remote machine through
SQL*Net. An Oracle instance is not present on the client machine.

This section provides a brief description of the most common pitfalls and problems encountered in a networked
environment.
Troubleshooting client server, and exploring SQL*Net in detail is beyond the scope of this manual. Please refer
to the SQL*Net manual for further information.

The files listener.ora and tnsnames.ora are important for a networked Oracle environment. These files can be
found in the directory:

• defined by the environment variable $TNS_ADMIN

• $ORACLE_HOME/network/admin if $TNS_ADMIN is not set

• /etc if not found in the above directories

The listener.ora file 监听程序

The listener process is a daemon process that runs on a machine which wants to accept connections over the
network from remote machines.

There is just one listener process running on a machine, even if there are multiple Oracle instances on the
machine.

The listener process has a configuration file associated with it. The configuration file, called the listener.ora file,
has information about all the instances on the local machine that are willing to accept networked connections,
and are willing to be serviced by the listener process. The following is a sample listener.ora file.

LISTENER=

(ADDRESS_LIST=

(ADDRESS=

(PROTOCOL=tcp)

(PORT=1521)

(HOST=minnie)

(ADDRESS=

(PROTOCOL=IPC)

(KEY=owminnie)

SID_LIST_LISTENER=
(SID_DESC=

(SID_NAME=owminnie)

(ORACLE_HOME=/export/home4/test/oracle)

The sample file services one instance, owminnie, on the host machine minnie. It accepts TCP and IPC
connections for the instance. The listener process listens on the reserved port number 1521 for incoming TCP
connections. The ORACLE_HOME variable associated with the instance owminnie is /export/home4/test/
oracle.

The installation process creates a listener.ora file for the instance you have installed.

单位机器上的listener.ora:

blade{owuser:/home/oracle/OraHome1/network/admin}% more listener.ora

# LISTENER.ORA Configuration File:/home/oracle/OraHome1/network/admin/listener.o

# Generated by Oracle configuration tools.

LISTENER =

(DESCRIPTION_LIST =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = blade.com)(PORT = 1521))

(DESCRIPTION =

(PROTOCOL_STACK =

(PRESENTATION = GIOP)
(SESSION = RAW)

(ADDRESS = (PROTOCOL = TCP)(HOST = blade.com)(PORT = 2481))

SID_LIST_LISTENER =

(SID_LIST =

(SID_DESC =

(SID_NAME = PLSExtProc)

(ORACLE_HOME = /home/oracle/OraHome1)

(PROGRAM = extproc)

(SID_DESC =

(GLOBAL_DBNAME = owblade)

(ORACLE_HOME = /home/oracle/OraHome1)

(SID_NAME = owblade)

blade{owuser:/home/oracle/OraHome1/network/admin}%

The tnsnames.ora file 范例列表文件

The tnsnames.ora file is a list of all the instances on the network that are available to the current instance,
including itself. The remote instances still reserve the right to restrict the current instance from connecting. A
sample tnsnames.ora file is as follows:
owminnie=

(DESCRIPTION=

(ADDRESS=

(PROTOCOL=tcp)

(HOST=minnie)

(PORT=1521)

(CONNECT_DATA=

(SID=owminnie)

owrocky=

(DESCRIPTION=

(ADDRESS=

(PROTOCOL=tcp)

(HOST=rocky)

(PORT=1521)

(CONNECT_DATA=

(SID=owrocky)

单位机器上的实例:

blade{owuser:/home/oracle/OraHome1/network/admin}% m tnsnames.ora

# TNSNAMES.ORA Configuration File:/home/oracle/OraHome1/network/admin/


tnsnames.ora
# Generated by Oracle configuration tools.

OWBLADE =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = blade.com)(PORT = 1521))

(CONNECT_DATA =

(SERVICE_NAME = owblade)

EXTPROC_CONNECTION_DATA.COM =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))

(CONNECT_DATA =

(SID = PLSExtProc)

(PRESENTATION = RO)

blade{owuser:/home/oracle/OraHome1/network/admin}%
With this tnsnames.ora the instance can connect to the instance owminnie, on the host minnie via the TCP
protocol. It can also connect to the instance owrocky on the host rocky via the TCP protocol. It is assumed that
the listener process is running on both those hosts and it listens on port 1521.

The tnsnames.ora file that is shipped with Oracle must be edited to work properly in an OpenWorks
environment. For example, formatting issues and the use of lower case letters in the file must be corrected so
that OpenWorks daemons will run.

Edit $ORACLE_HOME/network/admin/tnsnames.ora and make sure that the following words are all upper
case(保证为大写字母):

DESCRIPTION, ADDRESS, PROTOCOL, HOST, PORT,

CONNECT_DATA, SID

To troubleshoot problems with tnsnames.ora, run

$OWHOME/bin/owlistsids $ORACLE_HOME/network/admin/tnsnames.ora

Two values should be returned; the first is the ORACLE_SID, and the second is the host name. For example,

Owminnie minnie

单位机器上的的例子:

blade{owuser:/home/oracle/OraHome1/network/admin}% $OWHOME/bin/
owlistsids $ORACLE_HOME/network/admin/tnsnames.ora 回车

OWBLADE blade.com (显示结果)

blade{owuser:/home/oracle/OraHome1/network/admin}%

Troubleshooting Oracle Problems 排查Oracle数据库问题


This section describes the preliminary steps to take in troubleshooting the Oracle database system. For more
advanced troubleshooting techniques, refer to the appropriate section in the chapter “Oracle
Troubleshooting” in the Oracle Database Administrator’s Guide. If the problem cannot be corrected using
the techniques explained in any of

these sources, contact to your Landmark Technical Representative.

Database Processes May Not Be Running 数据库程序/进程可能未运行

If you suspect a problem with the Oracle database, you should check to make sure all processes are running.
Oracle has six essential database processes that must be running before you can access a database.
Process Name Purpose

Dispatcher ora_dnnn_<SID> An optional background process responsible for


routing requests from connected user processes
可选的后台调度程 to an available shared server process and
序 returning the responses back to the appropriate
user process.

Database Writer ora_dbwr_<SID> Performs all writes from the database cache in
memory to the database files on disk. Writes only
把内存中的数据库缓 occur when more data needs to be read into the
存写成磁盘数据库文 SGA and there is too little free space in the
件 database buffers.

Log Writer ora_lgwr_<SID> Performs all writes from the redo log buffers in
memory to the log files on disk. As transactions
把内存中的redo log commit, this process writes redo log entries into
写成磁盘log文件 an on-line redo log file.

Process Monitor ora_pmon_<SID> Watches the processes accessing Oracle and


makes sure that processes exit properly. Cleans
进程监督程序 up resources of processes that fail.

System Monitor ora_smon_<SID>


Ensures the database is ready for use when
Oracle is started. It performs instance recovery.
系统监督程序

Recovery Process ora_reco_<SID> Helps recover from failed distributed


transactions by rolling back the system (i.e.,
系统恢复程序 restoring the database to the condition it was in
before the transaction occurred).

These processes are started when the workstation is booted and should remain running until the workstation is
halted. Use the following command to verify that Oracle processes are running:

ps -efl | grep ora (IBM, Solaris, or SGI)

This command should display the following entries. (If you are running a client-only workstation, there are no
Oracle processes running on your workstation. Be sure Oracle is running on your server.)

409 0.0 1.5 228 1768 S 14:05 0:00 ora_pmon_ownova

410 0.0 2.0 200 2372 S 14:05 0:00 ora_dbwr_ownova


412 0.0 1.5 200 1776 S 14:05 0:00 ora_lgwr_ownova

413 0.0 0.0 468 0 IW 14:05 0:03 ora_smon_ownova

414 0.0 0.0 220 0 IW 14:05 0:00 ora_reco_ownova

415 0.0 0.0 200 136 IW 14:05 0:00 ora_s000_ownova

416 0.0 0.0 272 0 IW 14:05 0:00 ora_d000_ownova

If some of the processes are not running, restart the Oracle database server. To restart Oracle, exit from all
application programs and execute the following commands:

su - oracle

dbstart

lsnrctl start

If an error occurs during startup, it is displayed on the screen and logged in the file $ORACLE_HOME/rdbms/
log/alert_sid.log. If necessary, refer to the Oracle documentation for more detailed descriptions of log files and
error messages. The error message should indicate which processes did not start.

单位机器上的实例(该log文件会变得越来越大,可以人为将其减小):

blade{owuser:/home/oracle/OraHome1/rdbms/log}% ls

alert_owblade.log owblade_ora_1778.trc owblade_ora_2587.trc owblade_ora_3584.trc


owblade_ora_5626.trc

alert_owblade.log.old owblade_ora_2354.trc owblade_ora_2655.trc owblade_ora_3710.trc


owblade_ora_5639.trc

owblade_lgwr_558.trc owblade_ora_2396.trc owblade_ora_2935.trc owblade_ora_3826.trc


owblade_ora_6221.trc

owblade_lgwr_5652.trc owblade_ora_2453.trc owblade_ora_2949.trc owblade_ora_3862.trc


owblade_ora_7363.trc

owblade_ora_1142.trc owblade_ora_2467.trc owblade_ora_3102.trc owblade_ora_4070.trc

owblade_ora_1286.trc owblade_ora_2565.trc owblade_ora_3366.trc owblade_ora_5348.trc

blade{owuser:/home/oracle/OraHome1/rdbms/log}%
Unable to Connect to a Local Database 连不上本地数据库

Connecting to a local database requires that the Oracle processes are running (refer to the section “Database
Processes May Not Be Running” on page 85 for more information) and that you have authorization to access
the database. The utility program orauser lets

you provide user access to Oracle. For more details see the Manging the System manual and the OpenWorks
installation procedures.

Unable to Connect to a Remote Database 连不上远程数据库

Connecting to a remote database requires that the Oracle processes are running on the local and the remote
workstation and that you have permission to access the remote database. For information on database processes,
see “Database Processes May Not Be Running” on page 85.

For more details see the Manging the System manual and the OpenWorks installation procedures.

Unable to Halt the Oracle Processes 无法终止Oracle进程

The Oracle command dbshut is used to halt all Oracle processes before rebooting the workstation. You must
exit from all application programs before running this command. The dbshut command will not shutdown the
database if applications are still connected.(执行dbshut之前,要关闭OpenWorks应用程序)

Cannot Shut Down Database 无法关闭数据库

In many cases the shutdown normal and even the shutdown immediate commands will hang up the system. At
this point you have two options.

• Identify the processes connected to the system that are causing the hangup.

• Break or kill the shutdown process.

Identifying the current processes will have to be done through Server Manager, since no further standard
connection can occur during a shutdown. Breaking the process can work, but usually this results in the need for
a shutdown abort shortly after. The key is to make sure all connections are cleared before shutdown. If you do
not have that luxury, a shutdown immediate is worth trying. As a last resort, the abort command may be
unavoidable.

You may want to review the $ORACLE_HOME/rdbms/log/alter*.log files and any *_PID.trc files to see if
information has been written recently. Background processes can die or fail and the database writes these results
to the log and trace files. If these trace files do not provide immediate clues, a shutdown abort must be
performed. You have the

option of calling for support prior to shutting down, but it depends on your immediate needs.

Once the shutdown abort completes, immediately try to restart the system before doing any maintenance work,
unless you know it is absolutely necessary. If maintenance was scheduled, you can start with more confidence
after a clean shutdown.

Cannot Start Database 无法起动数据库

Usually, you only have to start the database if you have been forced to shut down or if the system requires
maintenance. This introduces a wide variety of problems of different severity. Once again the key is reviewing
the trace and log files to see if anything out of the ordinary has occurred during the last shutdown or during
system maintenance.(查看log文件,以确定问题出在何处)

A message such as:

ORA-01157 cannot identify file id - file not found

implies that the files listed in the control file cannot be accessed by the database during startup. If you do not
have the list of all files located within the database, you should start Server Manager using svrmgrl and issue the
connect internal command. Then spool the results from querying v$datafile and v$logfile with the database
startup mount command completed. These internal tables give you a full listing of the files used by the control
file (except for the control files themselves, which are listed in the config.ora file).

Files Moved or Changed 被移动或修改的文件

A file may have been moved, renamed, or deleted during standard operation or during shutdown. Most of these
cases can be corrected, but the file in question must be identified. Make sure that all of these files exist. If they do
not, you must carefully evaluate what has happened to the file in question. This is critical to your ability to
recover your current database without being trapped into recovering from the last week’s backup.

If any of the files were moved during the operation of the database, you have an even chance to recover the
current system without corruption.

If they were moved while the database was being shutdown, recovery is almost certain.

Recovery Strategies 数据恢复策略

There are two key factors to recovering your database in these cases.

• Is your system in archive mode? Is it archiving all online redo logs?

• Has the system been changed since the file was moved?

If the first case applies, this problem is beyond the scope of the current troubleshooting procedures, but can be
handled using Oracle techniques documented in the Oracle8 Server Administrator’s Guide.

If the second case applies, consider your situation carefully before proceeding:

1. Try to copy the most recent version of the datafile to the location required.

2. If the startup fails again, decide if the data is required to continue all other operations. Unfortunately if it is the
****** or RBS tablespace data files, you must return to your cold backup. If any other datafile is in question you
can simply drop the datafile before

the database starts up.

3. If you choose to drop the object before the database starts up, you are committing to the current database and
leaving all data and work associated with that file behind. Once the database is running, you can then re-create
the file if needed. The statements

required to perform this operation are clearly documented in the Understanding Oracle Guide.

Object or Data Lost 丢失的对象或数据

Typically, unless the environment has very tight security, a user will drop a table or remove all the rows from it.
In many cases this situation is unrecoverable. However, if the object was from the project tablespace, you can
recover the tablespace objects and data using the recover project function in the application. An export file is the
only

true means for recovering lost data or a dropped table.

WARNING: Beware of inconsistent data


Be aware that the export file may contain data that is inconsistent with the
current data in the project. Recovering just the one object lost from that export file is
possible but should be done with the aid of Technical Support to determine the
effects of that approach.

Database Resource Limits 数据库资源限制

While running an application, your process may pause, hang or fail with or without error messages. In most
cases, the failure of your process has resulted from a connection failure mentioned above.

However some other process limits may have caused the problem. You may see error messages such as:

ORA-01547 failed to allocate extent of size ‘num’ in

tablespace ‘name’

ORA-01562 failed to extend rollback segment (ID=num)

ORA-01631 max # of extents (num) reached in table name

These types of errors all stem from the resource availability within the Oracle database. In most cases where
resources are concerned, the user’s process will fail and the error message will be returned.

At this stage, you must assume that if you try the same process the same results will occur. Document the error
and then try the same process if the event was not very time-consuming. If you receive a consistent result you
should report the object in question to technical support for specific instructions on correcting the resources to
allow completion of the transaction.

Database Connection Problems 数据库连接问题


Your application connects to the system frequently when working with projects or application data. During this
process you can get any of the following errors:

ORA-01017: Invalid username/password.


ORA-01034:
Oracle not available.

ORA-12154: TNS Unable to connect to destination.

ORA-010145: User lacks create session privilege, logon denied.

There are many reasons these errors can occur but the checklist below is the quickest path to determining the
problem. With each item in the list there is a brief description of the tasks required to complete the checklist. If
you need more detail please refer to the Open Works System Administration Guide.

• Check your environment variables(检查环境变量). Oracle software and your access to it depend heavily
on the variables ORACLE_SID, ORACLE_HOME, and LD_LIBRARY_PATH to locate the software and
libraries required for Oracle products and applications. There are others for your particular application but
without these properly set, you cannot use the database.

• Check for database activity(检查数据库活动性). Log on using the Oracle account and use the following
command to check for background processes.

ps -al | grep ora (IBM, Solaris, or SGI)

You can confirm by logging into SQL*Plus directly using the following command:

sqlplus system/password

If the database is not running, review the log files in $ORACLE_HOME/rdbms/log/alter*.log to determine if
there were any problems with the database. Restart the system using the following commands to initialize the
Oracle background processes.

svrmgrl

connect internal

startup

• Check for network processes(检查网络进程). Log on to your workstation as oracle and verify that the
Listener process for SQL*Net is active.

This is an Oracle background process that connects users to the database. You can use the following command
to confirm the process is active.

lsnrctl stat
You can also confirm by connecting directly to the database. For example:

sqlplus system/password@ownova

If the network listener is not active, use the following command to activate the process.

lsnrctl start

• Check Oracle privileges for the user account(检查用户的Oracle权限). Each user account must have
sufficient privileges to connect to and access the database. To confirm this, use the following command:

sqlplus system/password

then review the information in the DBA_ROLE_PRIVS table to determine that the user has connect and
resource privileges. If not you can alter the user by entering the following SQL command:

grant ‘rolename’ to username;

Common Error Messages常见错误信息


Introduction 引言
Error messages are an important source of information to help you learn the cause of a problem. Error messages
could appear in any of the following locations:

• popup dialog boxes

• the console window

• the SeisWorks Main Menu window

• an error log file (s3d.err for example) in $HOME/run

• the xterm window from which you started an application

To insure that you see relevant error messages, keep both the console window and the main menu of the
application you are using open and visible as you work.

What’s in this Chapter本章内容


This chapter contains the following sections:

• Setting the Error Level on page 94(设置错误级别). This section describes the OpenWorks error
reporting system, which you can use to determine the causes of system problems.

• Using the OpenWorks Error Logger on page 99(使用错误信息日志). This section contains the
procedure for setting the error level using the OpenWorks Error Logger utility.

• Common OpenWorks Error Messages on page 100(OpenWorks常见错误信息). This section

describes typical OpenWorks error message and their solutions.

• Common SeisWorks Error Messages on page 103(SeisWorks常见错误信息). This section describes


typical SeisWorks error messages and their solutions.

Setting the Error Level 设置错误级别


The error level determines how severe a problem must be before you get an error message. The lower the error
level is, the more frequently you will receive error messages.

Error Level Types of Messages

4 No errors reported for OpenWorks programs.

3 默认级别 Severe or fatal errors. 严重或致命错误

2 Potentially serious errors.

1 Warnings.

0 Full history of all OpenWorks routines called.

The default setting is Error Level 3. Error Level 0 may provide too much information to be helpful.

For example, if you are having problems with SeisWorks/3D, you might lower the error level within an xterm
window, start SeisWorks/ 3D from that window, and repeat the sequence of steps that caused the problem in
your original session. This time, more detailed diagnostic information will be posted in the xterm window.

ERR_LEVEL controls error reporting.

ERR_LEVEL is an environmental variable that controls the level of error


reporting. It is typically set in the configuration file (lgcenv.cf) but could be set in
the initialization file (.lgcprofile or .lgclogin). See the OpenWorks System
Administration: Managing the System manual for information on locating and
setting environmental variables.

Checking the Error Level 检查错误级别


To determine the current error level, type

lgc_getenv ERR_LEVEL

Changing the Error Level 更改错误级别

For most diagnostic purposes, an error level of 2 provides sufficient information. The required commands are
slightly different if you are working in a Bourne shell or a C shell. The following examples show changing the
error level to diagnose SeisWorks problems.

In a Bourne Shell

To change the error level to 2, type the following commands:

ERR_LEVEL=2

export ERR_LEVEL

Now start SeisWorks from that same shell by typing

s2d (to start SeisWorks/2D)

or

s3d (to start SeisWorks/3D)

The Bourne shell posts detailed error messages as you work.

In a C Shell

To change the error level to 2, type

setenv ERR_LEVEL 2

Now start SeisWorks from that same shell by typing

s2d (to start SeisWorks/2D)

or

s3d (to start SeisWorks/3D)

The C shell posts detailed error messages as you work.

To Change the Error Level Globally 更改全局错误级别


The procedures outlined above for setting an error level within a Bourne or C shell
affect only those applications started from that shell. Applications started from other
windows or from the OpenWorks Command Menu will continue to use the error
level defined in the configuration file (lgcenv.cf) or initialization file (.lgcprofile or .
lgclogin).

If you wish to change the error level setting for all applications, you must edit the file
containing the ERR_LEVEL variable. See WHERE for instructions.

Viewing the Error Messages 查看错误信息

How you start the application influences where error messages are posted.

For Applications Started from the Command Menu 从命令菜单起动程序

When you start Landmark applications from the OpenWorks Command Menu, error messages are written to
the “run” directory. Typically run resides in your home directory.

Each application has its own “.err” file in the run directory. To see the error messages for SeisWorks, perform
the following steps:

1. Change to your home directory, then to the run directory by typing

cd

cd run

2. Display the SeisWorks error file by typing

more s2d.err (SeisWorks/2D)

or

more s3d.err (SeisWorks/3D)

3. To display and continually update the error file as you work, type

tail -f s2d.err (SeisWorks/2D)

or

tail -f s3d.err (SeisWorks/3D)

The file is overwritten when you open a new SeisWorks session.

For Applications Started from an xterm 从xterm起动的应用程序


When you start a Landmark application by entering its startup command in an xterm window, error messages
are automatically and continually posted in that window. You can direct those messages to a file. This frees the
window for other uses and captures the messages so that you can review or print them. There are two alternatives
for this method:

• directing error messages to a “script” file

• directing error messages to a text file

Directing Error Messages to a ‘script’ File 把错误信息定向到“角本”文件

This method records the error messages to a file you specify with the “script” command, as follows:

1. In an xterm window, set the error level to zero:

ERR_LEVEL=0 (in a Bourne shell)

export ERR_LEVEL

or

setenv ERR_LEVEL 0 (in a C shell)

2. Enter the “script” command:

script <myerrorfile>

3. Start your application in the same xterm using its executable name.

You can find these names in the launcher.dat or swlauncher.dat file, for instance. Each executable command
appears in the column after the application’s name; use only the first word. For example:

s2d (to start SeisWorks/2D)

or

s3d (to start SeisWorks/3D)

4. To view the resulting file, type

more <myerrorfile>

As the application runs, any errors display in the xterm and a copy is sent to the file. In addition, you can type
annotations of your workflow in the xterm, and this is also saved. When you finish running the application, exit
from the application menu. Then, in the xterm window type “exit” to exit from the script file.

Directing Error Messages to a Text File 把错误信息定向到文本文件


This method, which is similar to the preceding one, also sends error messages to a file. This method, however,
does not allow you to annotate your workflow in the file. To direct the messages to a text file, use one of the
following commands, depending on the type of shell you are in:

Bourne Shell:

s2d 1> <filename> 2>&1 (2D)

or

s3d 1> <filename> 2>&1 (3D)

C Shell:

s2d >& <filename> (2D)

or

s3d >& <filename> (3D)

To view the resulting file, type

more <filename>

Using the OpenWorks Error Logger 使用OpenWorks错误信息记录器


The OpenWorks Error Logger provides an easy way to set the error level for many OpenWorks utilities and for
StratWorks. However, it does not set the error level for SeisWorks.

To use the OpenWorks Error Logger, perform the following steps.

1. From the OpenWorks Command Menu, select System à Error Logger.

The following dialog box appears(从OpenWorks命令菜单上调用。图略).

2. Select the OpenWorks utility or application from the list on the left.

3. Click on the radio button for the Error Level you wish to use while running this application.

4. In the field labeled Error Log File, enter the full pathname of the file where you want the error messages
placed.

5. Select Controls à Run to start the application.

Error levels are same for the Error Logger and the keyboard.
The error levels used by the OpenWorks Error Logger are identical to those you
enter from the keyboard. 错误记录器上设置的错误级别与从键盘上设置的
级别是一致的。

Common OpenWorks Error Messages OpenWorks常见错误信息


The following section describes some common OpenWorks error messages and their solutions.

“The master database owsys could not be connected”连不上master数据库owsys

“Error reading the OpenWorks system database (owsys)!”读owsys时出错

Cause: Oracle is not running.

Solution: Check that the Oracle server is running. If it is not, restart the server by logging into the Oracle server
and issuing the following commands:

dbshut

dbstart

Cause: Your system is a client that is not mounted to the Oracle server.

Solution: In an xterm, check the OWSYSSID variable by typing:

echo $OWSYSSID

This variable identifies the host ID of the server where Oracle database is installed. Note that, by Landmark
convention, the OWSYSSID specification often includes an “ow” prefix. For example, if OWSYSSID is
specified as “owjazzman,” the actual host ID is “jazzman.”

Use the df command to determine whether your system is mounted to the Oracle server.

Cause: The OpenWorks variable OWSYSSID is not set correctly.

Solution: In an xterm, check the OWSYSSID variable by typing:

echo $OWSYSSID

If this variable is not set to the host ID of the system where Oracle is installed, use vi to edit the appropriate file.
Prefix the two letters “ow” to the hostname. Then exit from X, log out, and log back in.

Cause: The OpenWorks variable ORACLE_HOME is not set correctly.

Solution: Check the ORACLE_HOME variable by typing

echo $ORACLE_HOME
in an xterm window or by checking the Environment Status Tool.

实例:

blade{owuser:/home/blade/owuser}% echo $ORACLE_HOME

/home/oracle/OraHome1

blade{owuser:/home/blade/owuser}%

If this variable is not set to the directory where the Oracle database has been installed, use vi to edit the
appropriate Unix initialization file. Then exit from X, log out, and log back in.

“Unable to connect to the default $ORACLE_SID database”连不上缺省的


$ORACLE_SID数据库

Cause: You cannot reach the Oracle server because of a network problem.

Solution: Check with your system administrator to see if you have a network problem.

Cause: Your ORACLE_SID variable is not set correctly.

Solution: Check the System Configuration panel of the Environment Status Tool to see if ORACLE_SID is set
to the proper instance ID. (See “Tools for Troubleshooting” on page 17.)

If ORACLE_SID is not set, set it in the appropriate Unix configuration file.

“Cannot open the database instance configuration file”打不开数据库配置文件

Cause: You are trying to use the Database Sanity Checker to query the database instance configuration file from
a client.

Solution: The database instance configuration file can not be queried from a client. If you want to check the
setup of the Oracle server, you must do so from the server itself.

Cause: Your ORACLE_SID is not set correctly.

Solution: Check the System Configuration panel of the Environment Status Tool to see if the ORACLE_SID is
set to the proper Oracle server ID. If it is not, reset it in the appropriate Unix configuration file. (See “Tools for
Troubleshooting” on page 17.)

Cause: Your database instance configuration file does not have read permission. (Relevant only if your system is
an Oracle server.)

Solution: Generally, the database instance configuration file is called $ORACLE_HOME/dbs/config.ora. This
file should be readable by all. If the file does not currently have read permission, have your Oracle database
administrator change it so that it does.

“OpenWorks Database Project *nowell* Initialized”“nowell”初始化问题

Cause: Your seismic project is not associated with an OpenWorks project.

Solution: Ensure that your seismic project is associated with an OpenWorks project.

1. In an xterm window, type plist.

You receive a list of all currently available SeisWorks projects and their corresponding OpenWorks projects.

This list duplicates the contents of the plist.dat file.

2. Confirm that an OpenWorks project is listed against your seismic project.

3. If plist does not show a project associated with your seismic project, use the Seismic Project Associate utility
(available from the Seismic Project Manager’s Project menu) to associate an OpenWorks project with your
seismic project.

SeisWorks/2D users can find more information in the “Master Projects” and “Working Projects” chapters
of the SeisWorks/2D Project Management manual. SeisWorks/3D users should refer to the “3D Projects”
chapter of SeisWorks/3D Project Management.

An OpenWorks project is always necessary.

You cannot run SeisWorks without initializing an OpenWorks


project. You must always associate an OpenWorks project with a
SeisWorks project when you create the SeisWorks project. 主工区
必须与地震工区关联

“OpenWorks Database Project ‘<Project_Name>’: Access Failed”数据库的


<Project_Name>存取失败

Cause: OWSYSSID is not set correctly.

Solution: In an xterm, check the OWSYSSID variable by typing

echo $OWSYSSID

If this variable is not set to the host ID of the system where Oracle is installed, use vi to edit the appropriate file.
Prefix the two letters “ow” to the hostname. Then exit from X, log out, and log back in.

“You must be the Project Owner to Use this.”必须拥有工区所有者权限


Cause: You do not have authorization to modify, delete, or back up the selected project.

Solution: Have a user with MANAGE authority and the OW_ADMINISTRATOR role grant confer the same
authority and role to your account.

Common SeisWorks Error Messages SeisWorks常见错误信息


The following section describes some common SeisWorks error messages and their solutions.

“Cannot open .sm files. Cannot continue.” (SeisWorks/2D) 打不开.sm文件。无法继


续。(SeisWorks/2D)

Cause: The working project has been modified without updating the seismic map (.sm) files工作工区已更新,
但未更新地震图(.sm)文件.

Solution: Update the working project.

1. Exit SeisWorks.

2. At system prompt, type

update2d <projectname>

3. Reopen SeisWorks.

(注:该错误常见到。以后可按此法解决之。)

“Cannot set project” or “Cannot open project definition file”无法设置或打开工区定


义文件

SeisWorks/2D Solutions 解决办法

Cause: The application cannot open the project definition file (.ps2) for some reason.

Solution: Check permissions on the project definition file (.ps2), and change them to read-write if necessary.

(实例:

-rw-rw-rw- 1 owuser staff 5864 May 8 10:26 hjq072d.ps2 )

Cause: The application cannot find your project sys directory, and the project definition file (.ps2) resides there.

Solution: Check that dir.dat has been defined correctly. The location of this file is defined by $OW_PMPATH,
which is usually set to the directory $OWHOME/conf. However, if your system is running as a client or has
been modified at your site, dir.dat may be in $OW_DDF.

SeisWorks/3D Solutions 解决办法


Cause: The application cannot open the project definition file (.pds) for some reason.

Solution: Check permissions on the project definition file (.pds). Change them to read-write if necessary.

Cause: The application cannot find your project sys directory, and the project definition file (.pds) resides there.

Solution: Check that dir.dat has been defined correctly. This file usually resides in the $OW_PMPATH which is
often $OWHOME/conf. However, if your system is running as a client or has been customized, dir.dat may be
defined by $OW_DDF.

Common Solutions 常规解决办法

Cause: The application cannot open the project directories.

Solution: Check permissions on those directories. Change them to read-write-execute if necessary.

Cause: The environment is not properly defined.

Solution: Check that the environmental variable SEIS_DPATH is set. See VARIABLES for information on
variables.

Cause: You tried to open a session before selecting a project.

Solution: Select the project first, then open a new or existing session.

Cause: Someone has deleted the selected project from your system.

Solution: Restore the project using the SeisWorks Backup/Restore utility, or copy the project directories from
another system onto yours.(这是典型的工区恢复方法)

“Error in survey parameters in “.pdf” and “.hrz”观测参数错

Cause: The project as defined by master grid does not encompass the location or extent of your horizons.

Solution: Delete the horizon index file (.hrz)(层位索引文件) and recreate it using the HrzUtil utility. If this
does not solve the problem, call Landmark Customer Support.

“No fault segment/heave selected for assignment.”未选择断层段/高就进行断层关联

Cause: In trying to correlate faults, you selected a fault plane name without first selecting a fault segment or
heave.

Solution: Select a fault segment or heave, then select the fault plane you want to assign it to.

“.pd2 file not present” (SeisWorks/2D) “未出现.pd2文件”

Cause: You are trying to open a 3D project within SeisWorks/2D.


Solution: Exit SeisWorks/2D. Start SeisWorks/3D and choose the project again. Or, if you specified that project
by mistake, start SeisWorks/2D and choose the correct project.

“Viewport Not Allocated”未分配视口

Cause: The SEIS_DPATH environmental variable has not been properly set.

Solution: Set the SEIS_DPATH environmental variable in lgcenv.cf to $OWHOME/SeisUtils/dat

Give the absolute path for OWHOME. (Generally, lgcenv.cf is located in $OWHOME/conf.) (例:在
$OWHOME/conf 中,设置:SEIS_DPATH /home/OpenWorks/SeisUtils/dat )

“Virtual Graphics Error”权限问题

Cause: This often happens when you add a disk and create a directory with root permissions.

Solution: To give all users full access to all project directories, perform the following steps:

1. To assume superuser status, type

su

2. Now, as root, type

chmod -R 777 /p?

The above instructions assume, of course, that all your project directories begin with the letter p (/pa, /pb, /pc,
etc.).

If this is not the case, use the appropriate combination of letters and wildcards to include all the project
directories when you change the permissions.