On 8/23/2011 2:33 PM, Henrik Størner wrote:
On 23-08-2011 23:02, John Aldridge wrote:
I've been watching disk usage grow on a Xymon server for about the past six months, and this growth seems fairly steady and linear.
The big directories are histlogs and hostdata.
After digging into the documentation, I'm wondering what strategies Xymon admins prefer for managing disk usage.
hostdata: Used primarily for troubleshooting incidents. So you probably don't need to save this for very long. I keep it for one month, and simply run a cron-job to delete any files older than 30 days.
histlogs: Holds the detailed statuslogs that are linked to from the "History" and reports-pages. On the 1st of each month, I run a script to archive all of the history logs generated during the past month into a compressed cpio archive. They compress very well, so this is an effective way of archiving them. Then I delete all of the files that are older than 3 months.
With bash + GNU date, this will do it (run it on the 1st of the month):
Perhaps a patch to read <timestamp>.gz files in this dir as well and self pipe through gunzip on opening would be a nice feature?
That way you've still got live access and people compressing the files can leave them in place.
Eventually, it would be nice for xymon to use a compression library in other places too. For some links, it might be worth it to compress .data messages before sending them over, for example.
-jc