Troubleshooting
Troubleshooting
Agent halts abnormally, leaving ids_* files and message queues
❏If a running agent was not halted as described in “Halting
You should also remove any file in /var/opt/ids/ whose name starts with the string “ids_” and ends with a number (e.g., ids_1001). These are memory mapped files that are used by HIDS processes for interprocess communication. If they are not cleaned up, the corresponding partition might become full. A new memory mapped file will be automatically created the next time the agent starts a schedule. You should *not* remove any memory mapped files when a schedule is running.
|
| To clean up the IDS message queues |
Step | 1. | Determine which message queues are used by |
|
| ipcs |
Step | 2. | Remove the /var/opt/ids/ids_* files. |
Step | 3. | Remove each queue: |
|
| ipcrm |
Here’s an example of a hard kill followed by a message queue cleanup.
# ps | ids | # display the ids processes | |
ids | 16546 | 1 0 Apr 7 ? | 3:09 ./idsagent |
#kill
#ipcs
q | 602 0x000003e8 | ids | ids |
#ipcrm
Agent host appears to hang and/or you see message “disk full”
❏Check the local disk for available capacity. The following files have a tendency to become large and may need to be archived and truncated, or moved to a different disk partition with more space:
•/var/opt/ids/alert.log
•/var/opt/ids/gui/logs/hostname_alert.log
•/var/opt/ids/error.log
•/var/opt/ids/gui/logs/Trace.log
•/var/opt/ids/gui/guiError.log
Agent needs further troubleshooting
❏Create a directory for the logging information (for example, /var/log)
❏Restart the idsagent process with debugging enabled:
• /sbin/init.d/idsagent stop
242 | Appendix G |