This one time, at band camp, Eriberto said: > Hello. > > I delayed to write this message. Excuse me.
No problem, I'm a bit busy myself recently. > Stephen Gran escreveu: > > >This one time, at band camp, Eriberto said: > > > >>Hello! > >> > >>I set --max-children=20. However the clamav-milter died again. The > >>proccess runs 24h only. See the log in attach. > > > >I see nothing in the log indicating a problem, unfortunately. How > >familar are you with strace or ltrace? If you are familar with any of > >those, can you attach one of them to the running milter and send me the > >last couple of hundred lines after a crash. Note that the milter is > >threaded, so you'll want to run strace with -ff, and send the files > >relating to the last several threads before a crash. > > > I don't know strace or ltrace. I will try to use them. However I am in > vacance (30 days). I don't have contact with the systems. No problem - it can wait until you come back. > >gdb would also be nice - a stack trace after a crash would really help. > >If you're not familiar with any of these commands, let me know and I'll > >give you instructions. > > I observed the clamav without the -q and --quarantine-dir. It's works fine! You're saying it doesn't die when you don't use the --quarantine-dir option? That's really wierd, and smells like a leaking file descriptor bug or something. When you have time to debug this again, please run with --quarantine-dir, and take a look at the contents of the quarantine directory when it starts to choke. Are there tons of files? Is the disk full? What does an lsof | grep clamav-milter | wc -l say? Thanks, -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : [EMAIL PROTECTED] | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
signature.asc
Description: Digital signature