[pkg-fetchmail-maint] Bug#688015: memory leak

[Erik Thiele]

Package: fetchmail
Version: 6.3.18-2

/etc/fetchmailrc:

set daemon 30
set syslog
set no bouncemail
set no spambounce
set no softbounce

poll pop.myprovider.de proto pop3 user myusername at mypublicdomain.de pass
MYSECRETPASS smtphost mail.myinternaldomain.lan smtpname
myusername at myinternaldomain.lan fetchall

END OF /etc/fetchmailrc

the poll line above is repeated another 13 times with different usernames.

ps axuw|grep fetchmail

105        910  0.0 14.5  84756 74956 ?        Ss   Sep13   6:31
/usr/bin/fetchmail --bad-header accept -f /etc/fetchmailrc --pidfile
/var/run/fetchmail/fetchmail.pid --syslog

as can be seen, fetchmail is running as daemon. It has been manually
started on September 13, because the old instance was killed because of
kernel OOM (out of memory) kill. see:

Sep 12 07:03:50 mail kernel: [8986062.350113] fetchmail invoked
oom-killer: gfp_mask=0x201da, order=0, oom_adj=0
Sep 12 07:03:50 mail kernel: [8986062.350117] fetchmail cpuset=/
mems_allowed=0
Sep 12 07:03:50 mail kernel: [8986062.350119] Pid: 1444, comm: fetchmail
Not tainted 2.6.32-5-686 #1
Sep 12 07:03:50 mail kernel: [8986062.350121] Call Trace:
Sep 12 07:03:50 mail kernel: [8986062.350127]  [<c1089a78>] ?
oom_kill_process+0x60/0x201
Sep 12 07:03:50 mail kernel: [8986062.350130]  [<c1089ff5>] ?
__out_of_memory+0xf4/0x107
Sep 12 07:03:50 mail kernel: [8986062.350132]  [<c108a062>] ?
out_of_memory+0x5a/0x7c
Sep 12 07:03:50 mail kernel: [8986062.350135]  [<c108c90d>] ?
__alloc_pages_nodemask+0x3ef/0x4d9
Sep 12 07:03:50 mail kernel: [8986062.350138]  [<c108dced>] ?
__do_page_cache_readahead+0x98/0x16b
Sep 12 07:03:50 mail kernel: [8986062.350140]  [<c108ddd4>] ?
ra_submit+0x14/0x18
Sep 12 07:03:50 mail kernel: [8986062.350142]  [<c1088392>] ?
filemap_fault+0x16d/0x2e6
Sep 12 07:03:50 mail kernel: [8986062.350144]  [<c109a0b2>] ?
__do_fault+0x47/0x3b1
Sep 12 07:03:50 mail kernel: [8986062.350147]  [<c109c076>] ?
handle_mm_fault+0x48f/0x959
Sep 12 07:03:50 mail kernel: [8986062.350150]  [<c126dfa6>] ?
schedule+0x78f/0x7dc
Sep 12 07:03:50 mail kernel: [8986062.350153]  [<c1270d90>] ?
do_page_fault+0x2f1/0x307
Sep 12 07:03:50 mail kernel: [8986062.350155]  [<c1270a9f>] ?
do_page_fault+0x0/0x307
Sep 12 07:03:50 mail kernel: [8986062.350157]  [<c126f2f3>] ?
error_code+0x73/0x78

so I started a memory log:

2012-09-17 11,7% VIRT=63852 RES=60316 SHR=1888
2012-09-18 14,6% VIRT=84760 RES=74952 SHR=1884

so again it seems to run until oom death.

the machine is a virtual machine with 512MB ram + 1GB swap:

erik at mail:~$ free
             total       used       free     shared    buffers     cached
Mem:        514728     503600      11128          0      81100     237232
-/+ buffers/cache:     185268     329460
Swap:      1048568       2080    1046488


how can I further supply information on this issue? It is a production
machine, but maybe I can somehow help find the cause of the issue
anyway? Or is that memory leakage a known issue?


cya
Erik