qscand clamav process takes huge memory until server breaks

[herby4]

Hi,
I am using currently plesk 9.5.2 on Centos5 64bit in a VM. Now I decided to try the art packages and upgraded so I got php 5.2 and mysql updates.
Then I installed qmail-scanner and clamav.

For some hours everything is working perfectly, but then the system load goes up to 40.xx and all RAM + Swap is used until the Server is no more responding due to the qscand clamscan process as seen in top. Its forking like 10-15 times on process list.

Any hints?

br,
Herby

[scott]

It shouldnt use clamscan, it should be using clamdscan. If you didnt install clamd then thats probably whats going on

[herby4]

I installed together with all deps so clamd is also included:

Sep 17 11:06:25 yum: Installed: tnef-1.4.5-1.el5.art.x86_64
Sep 17 11:06:25 yum: Installed: 4:perl-suidperl-5.8.8-32.el5_5.1.x86_64
Sep 17 11:06:26 yum: Installed: maildrop-2.0.4-7.el5.art.x86_64
Sep 17 11:06:26 yum: Installed: daemontools-0.76-1.el5.art.x86_64
Sep 17 11:06:29 yum: Installed: clamav-db-0.96.2-1.el5.art.x86_64
Sep 17 11:06:39 yum: Installed: clamav-0.96.2-1.el5.art.x86_64
Sep 17 11:06:46 yum: Installed: 1:qmail-scanner-2.08-3.el5.art.x86_64
Sep 17 11:06:47 yum: Installed: clamd-0.96.2-1.el5.art.x86_64

clamd (PID 3643) running...

but top output gives clamscan:

Code: Select all

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 7591 qscand    22   0 82276  47m 2916 R 41.2  2.4   0:00.43 clamscan
 7593 qscand    21   0 73824  39m 2916 R 28.7  2.0   0:00.30 clamscan
...
...
3643 qscand    25   0  248m 195m 5944 S  0.0  9.9   0:05.81 clamd

in qmail-scanner.conf is:
# ClamAV settings
CLAMD_USER="qscand"

and clamd.conf is not touched.

br,
herby

[scott]

yup, you configured it wrong. Just run this: qmail-scanner-reconfigure

[herby4]

ok, will watch serverload now again.