Memory usage maxed and holding last few days

[jasonb3t]

The memory usage on our RHEL4 WHM server shot up from no more than 1.8GB to 8GB a few days ago and it's stayed up there.

Attached are images showing the trends for the last month and the last 24 hours or so.

The usual utilities (top, ps, etc.) don't show any processes using more than 130MB individually. Restarted many daemons individually with no significant drop in memory usage. Rebooted the server at 5am today and the usage was back up to the same level again within 1.5 hours.

Any ideas for causes or tools to investigate further?

 

[jasonb3t]

Possible lead

The only thing I can think of that might have coincided with this memory usage increase was that we ran out of space on the backup drive about when it started. We had a daily backup fail due to the space issue.

I've since deleted the monthly backup to make room and we've gotten a good daily backup again. Is it possible that the high memory is from cpbackup chunking data around at all hours? Will it go back to normal once the each stage of backup has stabilized?

 

[jasonb3t]

No drop in utilization yet

Many daily backups have completed and a weekly has completed. Still no drop in utilization yet.

 

[cPanelDon]

Any ideas for causes or tools to investigate further?

You may also try using "vmstat" to check memory usage information:

Examples:

# man vmstat
# vmstat -s -S M
# vmstat -S M 2 2
# vmstat -a -S M 2 2

To alleviate the issue one would need to first identify the specific software daemons, scripts, or processes that are utilizing the most memory, and then proceed from there to determine possible corrective measures if necessary.

I would also check if your OS may have been updated recently to help narrow the search for possible causes, such as a kernel update or other software updates via RPM; for CentOS/RHEL that use YUM as the package manager, you can check the following log file:

# less /var/log/yum.log