10-01-2010 12:15 AM
I'm seeing a 3-4 fold increase in the average CPU usage of nicitdl5.exe after server upgrade. OS is Server 2008 SP2 (x86). Half of clients are running 6.5 and others are running 6.2. I've had a crash also within a few days of install.
Database is ~3.5 GB. Lookout is running as service with delayed auto-start and nicitdl5.exe seems to need more time now to get squared away on restart so Lookout is not properly starting sometimes.
Database corruption? Mixed client issue? Other?
Thanks,
Ed
10-04-2010 03:50 AM
What's the number of the cpu usage?
Did you use server 2008 with 6.2 before?
The crash may be because of the database problem, corruption or something, especially when you are able to reproduce it every time you query a certain range of data.
Lookout server doesn't know the remote client's version. It's just the logos communication between them. The logos protocol should be compatible. So I don't think the mixed client can be the issue.
10-04-2010 12:20 PM - edited 10-04-2010 12:22 PM
Ryan,
CPU goes above 50% occasionally and spends a lot of time in the 20's and 30's. Normally, I see an occasional low teens number with most time spend in low to middle single digit percentages.
Yes, been with server 2008 since it's release.
Is the only way to repair a currupt database to experimentally archive sections of the database that do not cause MAX to crash? It would be really nice to be able to identify the corrupted time periods and/or traces somehow prior to archiving. It can take days to get most of the data archived succesfully, depending on how many and location of trouble spots.
Is there a reporting tool or health monitoring tool you use for this? Is there a log that Citadel makes of it's health? We need something to keep tabs on Citadel.... something besides crashes and/or CPU hogging.