Had a problem with VisionFS 3.0 on 5.0.5, started months ago, actually. Vision shares would be unavailable for 10 minutes or so every 5 hours or so. Found in the log (/usr/vision/vfsdata/logs/error.log) where VisionFS crashed with a Segment Violation. Searching for that brought memory errors, out of memory, etc. It was consistently periodic, every 4.5 to 5 hours, and would recover after about 10 minutes. (During that 10 minutes we cannot process credit card transactions at point-of-sale. Inconvenient.)
My solution was to add a cron job to simply restart VisionFS every hour, which is happening successfully.
New error on that: at one restart the log shows
2006/03/07 14:01:00.440 (pid 8382)
SCO VisionFS(3.0) ERROR:
locking/comms/failure
Failed to contact lock manager.
The server cannot run without the lock manager.
Check the vfslockd process is not being stopped.
Repeats every minute for 10 minutes, then there's a 15-minute period of nothing, then a successful start message for VisionFS.
I found this exact message with a solution for Unixware tuning, but could not correlate it to anything in OpenServer. They simply doubled existing values for MAXUP and SFNOLIM somewhere is system tuning. I have no Unixware experience, so I don't know what that means.
My solution was to add a cron job to simply restart VisionFS every hour, which is happening successfully.
New error on that: at one restart the log shows
2006/03/07 14:01:00.440 (pid 8382)
SCO VisionFS(3.0) ERROR:
locking/comms/failure
Failed to contact lock manager.
The server cannot run without the lock manager.
Check the vfslockd process is not being stopped.
Repeats every minute for 10 minutes, then there's a 15-minute period of nothing, then a successful start message for VisionFS.
I found this exact message with a solution for Unixware tuning, but could not correlate it to anything in OpenServer. They simply doubled existing values for MAXUP and SFNOLIM somewhere is system tuning. I have no Unixware experience, so I don't know what that means.