This one time, at band camp, Enrico Zini said: > although I purged freshclam (and clamav) from my system, a > "/usr/bin/freshclam -d --quiet" process was left running, which kept the > huge freshclam.log open, which left its disk space in /var/ allocated > although I had deleted the file manually. > > freshclam being purged meant that the stop script in /etc/init.d was not > there anymore to be used to shut freshclam down; killing the freshclam > process manually obviously put an end to all the sufferings and freed > that 1.2Gb I needed in my /var. > > I personally have no problems dealing with a scenario of "disk space is > not freed while the file is still open" coupled with "purging freshclam > leaves the daemon running" and "freshclam keeps its logfile open all the > time", but I can see how it all can cause a rather bad friday afternoon > to many, especiall because at that point "du" and friends will not show > you where the disk space is used anymore :(
du and df disagreeing is almost always a sign of an open but deleted file. lsof is your friend in these cases. It sounds like the main bug is that the init script failed to kill freshclam here. Can you provide any more info or help us to reproduce it? I know that freshclam blocks some of the useful shutdown signals while it's in the chunk of code concerned with downloads (which, without a network, will be more time than normal), but the initscript does try very hard to make sure it's killed off even so. Cheers, -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : sg...@debian.org | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
signature.asc
Description: Digital signature
_______________________________________________ Pkg-clamav-devel mailing list Pkg-clamav-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-clamav-devel
