Hello Glenn, Many thanks for the quick action! I'm very glad you still have savannah admin access :)
> On Feb 20, 2017, at 21:22, Glenn Morris <r...@gnu.org> wrote: > > And indeed the vcs0 /var partition was 100% full. > The size of /var was 3.8GB, and almost all of this was /var/cache/cgit. Sadly the entire disk on the new vcs0 VM is 9.8GB, so this was significant. Thanks for deleting the large files. > Obviously some limit needs to be put on the size of the cgit cache, > or this will happen again. Looks like /etc/cgitrc has cache-size=10000, > but apparently this refers to the number of entries. Does not help if > they are all 500MB... :( For now, I've reduced the number to 5000, and also added TTL on static pages cache to 5 minutes (the default was "never expire") and I suspect few of the large cache files were for what cgit calls "static" (=pages with fixed SHA1 in the URL). The new current settings in /etc/cgitrc are: cache-static-ttl=5 cache-size=5000 We might want to consider adding a cron job to delete old/large files ? something like: find /var/cache/cgit -type f -ctime +7 -size +2M -delete In the long run (as Bob wrote before), it would be better to have bigger disks, and perhaps put /var/ on a separate partition... one day... Thanks again, - assaf