 |
» |
|
|
|
|
|
|
|
This page lists errors you may encounter while using components of the AlphaServer Management Station (AMS). (See section D.4 of the AMS User Guide for advice on troubleshooting Event Manager
problems.)
- Problem: When a new ES47/ES80/GS1280 platform is added to the AMS/AMU on a Tru64 or OVMS management station, the SPM and AMU clients stop functioning after a few minutes. In
Tomcat's log file there are 'out of memory' errors followed by stack traces.
The Tomcat's Virtual Machine (VM), runs out of memory when AMU processes a platform whose MBM log in flash memory is full or almost full. This problem occurs on Tru64 and OVMS using the Fast VM. It
also occurs when using the Classic VM, but only when the initial (Xms) and maximum (Xmx) values for the memory allocation pool are not specified when starting Tomcat. Check the
installation documents for the Alpha Java SDKs for description of the VMs, memory management, and runtime options. To work around this issue we recommend the following:
Tru64 AMS/AMU installations
You can switch to Classic VM so that AMU can retrieve and save the MBM logs. After the logs are saved on the disk, you can switch back to Fast VM. Incremental increase of the logs with new entries do
not cause out of memory errors. Since the Classic VM is 4-5 times slower than the Fast VM, we recommend that you switch back.
To change to Classic VM perform the following:
#/sbin/init.d/amstomcat stop
#cd /usr/opt/ams/tomcat/bin
#cp setenv-classic.sh setenv.sh
#/sbin/init.d/amstomcat start
You need to check the AMU's log directory and verify that the log entries have been retrieved and saved into files on the disk. Using the classic VM, this operation will take about 1/2 hour to
complete. The AMU log files are located in /usr/opt/ams/tomcat/webapps/spm/WEB-INF/log directory. The filenames for the MBM logs have the format
Logs.MV-hex(nat IP).yyy-mmm-dd.log
AMU writes all the log entries that occurred the same date into one file. Once the log files are retrieved and populated, switch back to Fast VM. Note that once the logs are retrieved, any
additional entries thereafter are processed without any memory issues. You can now switch back to Fast VM.
#/sbin/init.d/amstomcat stop
#cd /usr/opt/ams/tomcat/bin
#cp setenv-fast.sh setenv.sh
#/sbin/init.d/amstomcat start
OpenVMS AMU installations
On OVMS the default VM is the Classic. To avoid out of memory errors caused by large MBM logs, it is only required that you set to set the Java runtime memory flags to large enough values. To
set the flags you need to create a data file that contain the VM flags, and then specify a symbol that points to this data file as follows:
$create sys$sysdevice:[amu]tomcat_jvm_args.dat
-ms128m -mx512m
^Z
$define /system apache$jakarta_java_parameters_file -
_$sys$sysdevice:[amu]tomcat_jvm_args.dat
To run the JVM with the memory range specified in the dat file above, you must also set the OVMS quotas according to the Java SDK documentation. You must run authorize and sysgen to increase the
quotas.
After you complete all the above, stop and re-start Tomcat.
$@sys$startup:apache$jakarta_shutdown
;
$@sys$startup:apache$jakarta_startup
- Problem: SPM displays "platform communication error" in the Hardware Warnings and Errors display when discovering a newly added platform or when refreshing an existing
platform.
The IP address of the NAT box or terminal server may be incorrect or the NAT box or terminal server may have not been configured correctly. Check the configurations, correct any errors you may find
and make sure that you can successfully send a ping command to the platforms before you try SPM again.
- Problem: The AMS software intermittently shows one or more platforms with a "Platform not responding" status. The AMU message pane may also show request time-out errors and a
partially drawn tree display.
Version 1.03x of the NAT box's firmware has tightened up access through its firewall (specifically in the Stateful Packet Inspection (SPI) parameters), and may shut down or limit communication via
the UDP protocol used by the AMS server to access the ES47/ES80/GS1280 SMLAN.
When the UDP packets are blocked, AMS times out waiting for a response from the AlphaServer. To resolve this problem, disable the SPI within the firewall function in the SMC7008ABR NAT box (8-port
only) with firmware v1.03x.
- Problem: On a Linux Red Hat AMS, the SPM client is launched successfully locally as an application as well as an applet using the web browser, but remote access using the
browser on a remote client machine fails.
You may have enabled the firewall when you installed or configured Red Hat. Reconfigure the firewall setting to No Firewall and try again.
Problem: In SPM, a partition's context menu does not contain a Telnet menu item
When a platform is added, the consoles that correspond to each partition need to be configured before they can be accessed using Telnet. After a console is successfully configured, the console's
context menu will contain "Telnet."
Problem: When trying to Telnet to a console using SPM or PCM, the console window displays the message "012 - Console is disabled."
Console access has been disabled by a user. In SPM, select the Enable Console menu of the console and make sure that it displays checked. In PCM, select the Enable check box in the console's
configuration window.
Problem: While using SPM a dialog with the message "The AMS server has stopped responding..." is displayed. Closing this dialog causes the SPM display to be cleared.
You may have had a network interruption and the SPM client could not communicate with the server. If your network is working and you can successfully send a ping command
to the server, then the Tomcat server may stopped running. Log into your AMS server and determine if the Tomcat process is still running by entering one of the following commands
On Tru64 UNIX systems:
# /sbin/init.d/amstomcat status
On Linux systems:
# /etc/init.d/amstomcat status
If the catalina process is not running, make sure that Tomcat was not stopped from the command line by checking the last line in /usr/opt/ams/tomcat/catalina.out. If that line is
"Stopping service Tomcat-Apache", it means that Tomcat was stopped by a user. To restart Tomcat, enter one of the following commands:
On Tru64 UNIX systems:
# /sbin/init.d/amstomcat start
On Linux systems:
# /etc/init.d/amstomcat start
If Tomcat was not stopped by a user, Tomcat and the JVM may have crashed. The following steps will help you collect debug information before you contact AMS support
- Stop Tomcat if it is running.
- Collect and save current log files from the following directories :
/usr/opt/ams/tomcat/logs
/usr/opt/ams/tomcat/webapps/spm/WEB-INF/log
/usr/opt/ams/logs (all files, including subdirectories
- Enter the following commands to enable the SPM and AMU server debug:
# cd /usr/opt/ams/tomcat/webapps/spm/WEB-INF
# cp debug.SPMSLog4J.xml SPMSLog4J.xml
# cd data
# cp debug.ServerConfiguration.xml ServerConfiguration.xml
- Start Tomcat and try to reproduce the problem.
- When the problem is reproduced, collect the log files as in step 2.
- Send the log files from steps 2 and 5 to AMS support with a description of the problem.
- Stop Tomcat.
- Copy the non-debug files back to stop collecting debug information:
# cd /usr/opt/ams/tomcat/webapps/spm/WEB-INF
# cp default.SPMSLog4J.xml SPMSLog4J.xml
# cd data
# cp default.ServerConfiguration.xml ServerConfiguration.xml
- Start Tomcat
Problem: PCM exits after entering a regular expression in the Events Add dialog box.
You may have entered double-quotes in the regular expression. Double-quotes cause the PCM to exit.
Problem: A blank screen displays when attempting to run the SPM.
The AMS machine may not have enough memory to run the SPM. See the installation instructions for more information.
Problem: Updating the platform's firmware is unsuccessful.
- Stop the Tomcat server before attempting to update the firmware by using one of the following commands:
For Tru64 UNIX:
# /sbin/init.d/amstomcat stop
For Linux:
# /etc/init.d/amstomcat stop
- Use the command line interface to the platform's management port by connecting to it from either the SPM or the PCM.
Problem: The Console Management Facility daemon, cmfd, keeps writing to console logs in /usr/opt/ams/logs/cmf.current and
/usr/opt/ams/logs/cmf.dated until the disk is full.
Copy the contents of the log files to a separate location and empty the contents of the log directories.
Problem: The PCM cannot connect to the console daemon, cmfd, and the console status is unknown. The PCM displays this error message in a dialog box and in the list of managed
systems. Restart the cmfd daemon.
Problem: When using the PCM, you attempt to connect to a console but the connection is closed by a foreign host and you are asked to press Return to return to the PCM main
window.
This may be caused by another user or application being connected to the console. Ask the user to disconnect from the console or clear the console line.
Problem: A console has a status of Inaccessible.
This may be caused by another application using the console connection or by a VT100 terminal that is connected directly to the port. Ask the user to disconnect the terminal from the console or clear
the console line.
Problem: A console has a status of Unknown.
This may occur if the cmfd daemon has stopped
or unresponsive. To check the status of the cmfd
daemon, enter one of the following commands:
On Tru64 UNIX systems:
# /sbin/init.d/cmfd status
On Linux systems:
# /etc/init.d/cmfd status
If the cmfd daemon is not running, enter one of the following commands to start it:
On Tru64 UNIX systems:
# /sbin/init.d/cmfd start
On Linux systems:
# /etc/init.d/cmfd start
|
|
|
