guam - high number of crashes

Valentin Laskov laskov at festa.bg
Tue Mar 10 16:26:52 CET 2020


Hi,

Recently I wrote about my troubles with guam. I think that the core of 
my problem is a slow storage while some users are migrating their mails 
to my server.

top shows very high percent (54.5 and more) waiting for a storage operations

%Cpu(s):  6,3 us,  2,7 sy,  0,0 ni, 36,5 id, 54,5 wa,  0,0 hi, 0,1 si,  
0,0 s


Also I supose there is a bug in guam or in erlang . top shows that 
beam.smp consumes more and more virtual memory.

   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
26108 root      20   0   13,0g   6,1g    908 S  21,1 79,7  23:57.30 beam.smp

When the virtual memory runs out the kernel kills guam and systemd 
starts it again. dmesg shows

Out of memory: Kill process 9123 (beam.smp) score 791 or sacrifice child
Killed process 9123 (beam.smp), UID 0, total-vm:21696728kB, 
anon-rss:5484380kB, file-rss:0kB, shmem-rss:0kB

I forced myself to discard syncroton log messages by

ln -s /dev/null /var/log/kolab-syncroton/console.log

and imapd log messages too

Regards!
Valentin

На 10.03.2020 в 16:37, David Obando написа:
> Hi,
>
> in my pretty fresh kolab (16.1) I have some trouble with a high number
> of guam crashes and suspect the crashes to also destabilize cyrus.
>
> I currently have 5 domains and 21 mailboxes in use, plan is to scale
> higher. Because of the occuring guam/cyrus issues I stopped migrating
> more users to the kolab system.
>
> I have thousands of crash reports per day, syslog then shows:
>
> Mar 10 12:02:26 mail guam[13456]: 12:02:26.753 [error] CRASH REPORT
> Process <0.17627.1> with 0 neighbours exited with reason: bad argument
> in call to
> erlang:bit_size([[<<"FwFA45zxjFeGa7qcQAEPKDOCRwuc9jxj0r0rxPdwmzeJ5CC3yN0xjt1\r\n98V8++MNVmigeaAKzA4jTnaG6AHB6Zr6LC...">>],...])
> in eimap:joined/2 line 487
>
> Logfiles in /var/log/guam show detailed CRASH and ERROR reports.
>
> Additionally I'll receive the following syslog errors for several
> existing mailboxes:
>
> Mar 10 15:23:49 mail cyrus/imaps[2577]: IDLE: error sending message INIT
> to idled for mailbox example.de!user.alice: No such file or directory.
> Falling back to polling every 60 seconds.
>
> Testusers describe errors when accessing their mailboxes with
> mailclients (e.g. Thunderbird), like mails aren't downloaded, after
> sending mails can't be stored in sent folder.
>
>
> Does anyone have a clue, what could be wrong or how I could fix my issues?
>
>
> Thanks very much!
>
> David
>
>
>
>


More information about the users mailing list