Description of problem: By default, it appears to extract gzipped archives into a working dir in /var/cache/logwatch/. Which has a high potential to blow up the /var volume if, for example, apache is a high-volume site and the logs are already rotated and zipped in order to prevent just such a problem. :-/ Even worse, when logwatch blows up /var and dies, it leaves all the temp junk behind so what would have been a brief blowup of space on /var instead is a permanent disk full error requiring manual intervention. Version-Release number of selected component (if applicable): This was verified on v7.0, but a brief inspection of the v7.2 update does not indicate that this issue was addressed. How reproducible: Roll out high-volume apache server with logwatch installed. Setup a logrotate policy as needed for faster rotation and compression. Watch closely at 04:04 logwatch consume growingly huge amounts of space as it expands the compressed logfiles. If it every fills up /var, BANG it will leave behind its mess and a full volume. Expected results: #1 required: it should die gracefully and clean up its mess #2 optional: some sort of sanity check before going to town on compressed huge archive files
Thanks. I discuss this problem with upstream (thanks Bjorn) the upstream solution is now built in the last version of logwatch - logwatch-7.3-2. Could you test this version? If there is any problem, please add your comment here.