On 2016-05-13 14:06:27 [-0600], Will Aoki wrote: > That's it. A space is logged after the colon, but nothing else.
interresting. The source says that there should be more. Your bug report says you run i386. Is this the case for the server or just the machine you made the report? > May 13 08:56:05 skunk clamd[12310]: accept() failed: > May 13 08:56:05 skunk clamd: Last message 'accept() failed: ' repeated 198 > times, suppressed by syslog-ng on [the loghost] > May 13 08:56:11 skunk clamd[12310]: accept() failed: > May 13 08:56:11 skunk clamd: Last message 'accept() failed: ' repeated 199 > times, suppressed by syslog-ng on [the loghost] Ah right, there two loops. So clamd tries mutliple times and somehow always runto into the same error. > I've noticed that /tmp is filling up with directories named e.g. > clamav-fe97224f9fa888d6e2d47ddfee0ca573.tmp Those are created by some filetypes during scan but should have been cleaned up. You can remove them. But this looks like it terminated in a hurry. Could you enable debug loglevel? Maybe it logs something usefull. Also would it be possible to enable core dumps and see if it dumps something? Could start clamd from commandline in foreground so you can log its exit code? Ah. And could please check with top or something if the memory of clamd grows overtime? In case it has a memory leak somewhere. If you manage to give me something to reproduce it locally (like a VM image) I might try it when I have some time. Sebastian
