I'm running amavis 2.3.2 on a redhat AS4 update 1
system. The MTA is postfix: 2.2.5.
My system runs amavis with max_servers=10
Problem is after several hours one or more amavis childs
hangs/loops.
[root@lxhv1m02 ~]# ps -ef | grep amavisd
vscan 1343 1 0 Nov07 ? 00:00:12 amavisd
(master)
vscan 29475 1343 2 Nov08 ? 00:31:37 amavisd
(ch9-finish)
vscan 6074 1343 2 Nov08 ? 00:26:22 amavisd
(ch6-finish)
vscan 17761 1343 1 10:09 ? 00:00:05 amavisd
(ch6-17761-06)
vscan 17764 1343 1 10:09 ? 00:00:04 amavisd
(ch8-17764-08-2)
vscan 17834 1343 2 10:11 ? 00:00:04 amavisd
(ch4-17834-04-2)
vscan 17914 1343 5 10:13 ? 00:00:02 amavisd
(ch1-17914-01-4)
vscan 17915 1343 9 10:13 ? 00:00:03 amavisd
(ch1-17915-01-4)
vscan 17922 1343 3 10:13 ? 00:00:00 amavisd
(ch1-17922-01-3)
vscan 17926 1343 5 10:13 ? 00:00:01 amavisd
(ch1-17926-01-2)
vscan 17930 1343 6 10:14 ? 00:00:01 amavisd
(ch1-17930-01-2)
root 17943 17427 0 10:14 pts/0 00:00:00 grep
amavisd
the avaliable memory goes down:
top - 10:13:50 up 60 days, 22:17, 4 users, load
average: 0.15, 0.34, 0.37
Tasks: 153 total, 1 running, 152 sleeping, 0
stopped, 0 zombie
Cpu(s): 8.5% us, 1.5% sy, 0.0% ni, 89.5% id, 0.0%
wa, 0.0% hi, 0.5% si
Mem: 2073896k total, 2039560k used, 34336k
free, 137220k buffers
Swap: 2048248k total, 189164k used, 1859084k
free, 117924k cached
strace shows me that the process is in a loop
If I don't kill this process amavisd (master) can't fork
and die. This is because the looping childs grows on
and each of them claims huge of mem.
If I kill this processes the system is running fine.
Best regards,
Ruediger Hofer
Ruediger.Hofer@leoni.com
phone 49-911-2023-220
strace from the loop process
Logged In: YES
user_id=28904
Originator: NO
Use the amavis-user mailing list for amavisd-new support.