General discussion on installation and configuration of SOGo

Text archives Help


Re: [SOGo] Can't start, "Default zone has run out of memory"


Chronological Thread 
  • From: "\" \"" ( ) < >
  • To:
  • Subject: Re: [SOGo] Can't start, "Default zone has run out of memory"
  • Date: Tue, 28 Feb 2017 14:13:30 +0100
  • Dmarc-filter: OpenDMARC Filter v1.2.0 mail.inverse.ca 3DCA81201A5E

Hi,

Yes, it's a small LAMP web hosting server. But none of the other applications has reserved RAM, for what I know. There's more than 1 GB free right now and I already tried increasing to 4 GB total with 3 GB free.
I also tried to know the memory size after reading GNUstep documentation but I can't get to use NSZoneStats in command line, I don't know how to use the function.
http://www.delorie.com/gnu/docs/gstep-core/gnustep-zones_15.html
I expected something like "defaults read sogod NSZoneStats" to work. No : "couldn't read default".


Michaël





On 28/02/17 13:31, Christian Mack
( )
wrote:
Hello

As zones are used to seperate different applications.
Is there another application running on that server, which has a lot of
RAM reserved to it?
How big is the RAM size for the "Default zone" ?


Kind regards,
Christian Mack

Am 27.02.2017 um 17:26 schrieb
" "

( ):
SxVMemLimit is per worker.
Yeah, it was some kind of desesperate try changing SxVMemLimit :/

How many workers do you try to start?
The issue occurs as soon I try to start a fresh Sogo reinstall, with the
default config. It's in use for a SOHO, 1-2 worker is enough.


Michaël




On 27/02/17 17:04, Christian Mack
( )
wrote:
Hello

Am 27.02.2017 um 10:50 schrieb
" "

( ):
Bonjour,

For a few years I had Sogo 2 running but this week-end I noticed
Thunderbird wasn't able to retrieve the cardbook. All suddenly, I don't
remember any relevant change on the server.

Sogo daemon is not running anymore. In /var/log/sogo/sogo.log I get this
: "/usr/sbin/sogod: Uncaught exception NSMallocException, reason:
Default zone has run out of memory". But a few lines above I read "vmem
size check enabled: shutting down app when vmem > 384 MB. Currently at
175 MB" so it seems good. I also tried to increase SxVMemLimit just in
case and the result is the same as soon I try to start the daemon.

Does it relates to GNUstep ?
The only clue I have is that the virtual host RAM was reduced from 2 to
1 GB lately before the issue occurs. But now I switched back to 2 GB.
The "free" command says >1 GB available free RAM.

I upgraded to Sogo 3, I purged Sogo and GNUstep then reinstalled, I
rebooted the system twice : nothing changed with a fresh reinstall.

I use Sogo 3 nightly build, Debian 7 Wheezy and Postgresql.

Thanks a lot for your help.

SxVMemLimit is per worker.
How many workers do you try to start?


Kind regards,
Christian Mack





Archive powered by MHonArc 2.6.18.

Top of page