-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Jay Deiman wrote: > This has been submitted as bug 1567: > https://wwws.clamav.net/bugzilla/show_bug.cgi?id=1567 > >
So, I have been working the above bug now for a while and the latest scenario involved me setting up a linux box so I can see if I experience the same memory leak as I've been seeing in FreeBSD. This has brought about a new, different issue. The issue I'm running into on this linux box (Gentoo, up to date with clamd 0.95.1) is that I end up with all my clam tcp STREAM connections sitting in a LAST_ACK state for an extended period of time, not quite sure how long exactly, but long enough that hit a limit of about 49XX connections in this state and clamd starts to spew "accept() failed" messages into the log file. Doing some digging, I found a bug from long ago where this was an issue with the exact same setup I have (clamav-milter on FreeBSD hitting clamd on a Gentoo linux machine) and it doesn't look like anything was actually done about it since it's marked as "RESOLVED WORKSFORME". https://wwws.clamav.net/bugzilla/show_bug.cgi?id=487 Has anyone else experienced this problem? I'm thinking I might try a different linux distro as a sanity check, but I'd like to know if anyone has any ideas or perhaps sysctl tweaks that I could try for this. To note, all the FreeBSD clamav-milter clients are running clamav-milter 0.94.2, not 0.95.1 and the FreeBSD versions range from 6.2 to 7.1. Jay Deiman -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkn3XJ4ACgkQQ0lr+ZVKSBi98ACcDF/LZcmY81goR/GRumJMixIh QmYAoIuMU399By3kiSdRd1hRhTMvi7X/ =F9RS -----END PGP SIGNATURE----- _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://www.clamav.net/support/ml
