On 27/01/12 12:52, Christian Naumer wrote:
We have the same problem. I was putting it down to the switch to 64bit
architecture. I increased SxVMemLimit to 1024 to no avail. The problem
occurred less with 1024 but was not gone.
Yes that makes sense. But I don't want to keep raising the mem unless
there's a good reason for it. We never had these issues a while back.
I'd rather try to find what is making SOGo collapse so easily when
previously with less memory it was so robust ....
We have an application that fetches a ical (large file>2000 entries)
file from sogo an this caused this issue every time. After I stopped
fetching this ical we don't have any more crashes with SxVMemLimit=1024
with less this still happens.
We AFAIK don't have anyone loading ical files. I'm still not totally
sure what is causing the excessive ram usage. Recently I did see one
worker getting killed, (first example in earlier email), when we had to
clear a Lightening cache and mem / cpu increased when Lightning was
loading all the events from the server anew : about a 1800kB cache file
(in Lightning). But this isnt that much really is it ? We only have 30
users, 5 calendars per user ..
Also when I try to export this calendar in the web interface the same
thing happens.
I posted this some time ago:
http://comments.gmane.org/gmane.comp.groupware.sogo.user/7774
Yes I think I remember that one, but didnt have any useful suggestions
Julian
--
[email protected]
https://inverse.ca/sogo/lists
