Page MenuHomePhabricator

zoomviewer is using 837G out of 8T for Tools
Closed, ResolvedPublic

Description

@dschwen almost all of this is in

836G public_html/cache

Is there any way we can clean this out as a one-time fix, if not periodically?

Information on maintainers taken from https://wikitech.wikimedia.org/w/index.php?title=Special:NovaServiceGroup&action=managemembers&projectname=tools&servicegroupname=tools.zoomviewer and then https://wikitech.wikimedia.org/wiki/User:Dschwen

Event Timeline

chasemp updated the task description. (Show Details)

Hey Chase, I went ahead and deleted cache entries older than 90 days.

find /data/project/zoomviewer/public_html/cache -mtime +90 -delete

I can put this in the tool's crontab, but I'd like to observe the growth rate of that directory for a few days. Maybe I should rather make the check on atime (if that is available on that filesystem).

@dschwen, thank you.

This tool just went from the #1 space user to somewhere around #20. Much appreciated. I believe our mounts are all done w/ noatime. Let me know what you decide for ongoing cleanup and if I can help.

Hi @dschwen, zoomviewer is currently using 408G/8T of tools NFS storage. Could you consider cleaning up old data, and also revisiting adding the cleanup command to the tools crontab? Thanks!

Done. I put the find command into a small script and added it to the crontab (via jsub).

scfc subscribed.

prune.sh (with 90 days -mtime) has been added to crontab, so considering this resolved. If this is still unbearable, it might be interesting to analyze access.log how often cached files are revisited, e. g. if 99 % of all files were visited only once, the cache age could probably be reduced drastically without impacting the average user.