1.4/X OS Station Lockup Debug/Troubleshooting Guide
- How To Debug an Operator Station
- Note what operator actions, if any, prompted this event. Note displays and dialogs that are active during this time. Examine system diagnostics at that time and note any system level activity, (database installs, etc.) that was taking place during that time. Identify if the station is locked up (frozen) or if it has crashed.
There are basically two kinds of lockups, the difference being keyboard input availability. If possible, before rebooting a station that is locked up, attempt to access the system performance page for that OS from another node on the system and note any abnormally high statistics.
The recommended way to reboot a station that is locked up is to remotely login as the 'root' user and issue a shutdown command. When this is done, the OS will create some log files that will be discussed later.
If the machine is simply powered down, these files do not get created. When a station crashes, some log files are created as well as a core file. Creating a core file may take on the order of minutes, so before a station is rebooted, attempt to determine if the disk light is continually active, (indicating the core file is being created). The station will attempt to restart itself after a crash. After a machine has crashed or has been rebooted due to a lockup, log files need to be examined or faxed in for analysis.
From a HP terminal window, go to the /products/log directory. Enter the 'll -rt' command, (this will sort files in reverse order).
At the time of the crash or reboot, there should be a file called osSummary.out if the system crashed, or osShutdown.log if the station was rebooted due to a lockup. These files have the same basic format. The most important section of these files is the section that looks for running tasks.
A list of expected tasks is presented and if a task is not running there will be, a '... Not Found' indication. All tasks not found should be noted.
- My Operator Station has locked up