subject:"\[SOGo\] Weird load issue"

Re: [SOGo] Weird load issue

2011-11-30 Thread Sebastian Reitenbach

 Hi,

On Tuesday, November 29, 2011 19:25 CET, Ludovic Marcotte 
lmarco...@inverse.ca wrote: 
 
 On 29/11/11 13:23, Donny Brooks wrote:
  The only reference to the WOWorkerCount is for the GNUStepDefaults 
  file. I am running Centos 5.4 currently on that machine and have no 
  /etc/sogo anything.
 And what's the value? What does ps aux  | grep sogod tells you?

also for me on OpenBSD, from time to time, sogod freaks out. One sogod process 
is running the Watchdog, and the other, is busy. This usually happens when in 
the web interface, reading mail. Connecting with gdb to the process, a stack 
trace looks like this:


Program received signal SIGINT, Interrupt.
[Switching to process 3442, thread 0x6baf7000]
0x5f5574a9 in _la (self=0x6514b8e8, _laCnt=0) at NGImap4ResponseParser.m:66
66register unsigned char c = __la(self, _laCnt);
(gdb) bt
#0  0x5f5574a9 in _la (self=0x6514b8e8, _laCnt=0) at NGImap4ResponseParser.m:66
#1  0x5f55766e in -[NGImap4ResponseParser parseResponseForTagId:exception:] 
(self=0x6514b8e8, _cmd=0x5f5d7b18, _tag=6, ex_=0xcfbe2944) at 
NGImap4ResponseParser.m:271
#2  0x5f5695a6 in -[NGImap4Client 
processCommand:withTag:withNotification:logText:] (self=0x6573c188, 
_cmd=0x5f5d7f50, _command=0x5f5d6c10, _tag=1 '\001', _notification=1 '\001', 
_txt=0x5f5d6c10) at NGImap4Client.m:1552
#3  0x5f569bf8 in -[NGImap4Client processCommand:] (self=0x6573c188, 
_cmd=0x5f5d7b38, _command=0x5f5d6c10) at NGImap4Client.m:1623
#4  0x5f564343 in -[NGImap4Client logout] (self=0x6573c188, _cmd=0x5f5d7aa8) at 
NGImap4Client.m:604
#5  0x5f5625a2 in -[NGImap4Client dealloc] (self=0x6573c188, _cmd=0x65c526d0) 
at NGImap4Client.m:232
#6  0x65a1301a in -[NSObject release] (self=0x6573c188, _cmd=0x5f5e15e0) at 
NSObject.m:1919
#7  0x5f59586e in -[NGImap4Connection dealloc] (self=0x5c052088, 
_cmd=0x65c526d0) at NGImap4Connection.m:92
#8  0x65a1301a in -[NSObject release] (self=0x5c052088, _cmd=0x5fc49bc8) at 
NSObject.m:1919
#9  0x5fc26c15 in -[SOGoMailBaseObject dealloc] (self=0x665023c8, 
_cmd=0x5fc4d908) at SOGoMailBaseObject.m:63
#10 0x5fc2c1f5 in -[SOGoMailFolder dealloc] (self=0x665023c8, _cmd=0x65c526d0) 
at SOGoMailFolder.m:139
#11 0x65a1301a in -[NSObject release] (self=0x665023c8, _cmd=0x65c0ec90) at 
NSObject.m:1919
#12 0x658f7bfd in -[GSArray dealloc] (self=0x63894f28, _cmd=0x65c526d0) at 
GSArray.m:136
#13 0x65a1301a in -[NSObject release] (self=0x63894f28, _cmd=0x65c0ed78) at 
NSObject.m:1919
#14 0x658fa26f in -[GSArrayEnumerator dealloc] (self=0x5c055108, 
_cmd=0x65c526d0) at GSArray.m:942
#15 0x65a1301a in -[NSObject release] (self=0x5c055108, _cmd=0x65c21998) at 
NSObject.m:1919
#16 0x65951685 in -[NSAutoreleasePool emptyPool] (self=0x6748c748, 
_cmd=0x65c21988) at NSAutoreleasePool.m:603
#17 0x65951455 in -[NSAutoreleasePool dealloc] (self=0x6748c748, 
_cmd=0x65c21980) at NSAutoreleasePool.m:485
#18 0x6595140a in -[NSAutoreleasePool release] (self=0x6748c748, 
_cmd=0x5dda4168) at NSAutoreleasePool.m:478
#19 0x5dcd2533 in -[WOHttpAdaptor acceptControlMessage:] (self=0x6748cfc8, 
_cmd=0x5dda4068, aNotification=0x692ed948) at WOHttpAdaptor.m:522
#20 0x65a019fb in -[NSNotificationCenter _postAndRelease:] (self=0x686e0f78, 
_cmd=0x65c4bf30, notification=0x692ed948) at NSNotificationCenter.m:1162
#21 0x65a02345 in -[NSNotificationCenter postNotificationName:object:userInfo:] 
(self=0x686e0f78, _cmd=0x65c4bf38, name=0x6a3fe160, object=0x6989bc08, 
info=0x0) at NSNotificationCenter.m:1221
#22 0x65a021dd in -[NSNotificationCenter postNotificationName:object:] 
(self=0x686e0f78, _cmd=0x6a3fe268, name=0x6a3fe160, object=0x6989bc08) at 
NSNotificationCenter.m:1201
#23 0x6a3c9c00 in -[NSObject(FileObjectWatcher) 
receivedEvent:type:extra:forMode:] (self=0x6989bc08, _cmd=0x65ca65d0, 
_fdData=0x8, _type=ET_RDESC, _extra=0x8, _mode=0x65c61b18) at 
NSRunLoop+FileObjects.m:57
#24 0x65b29b39 in -[GSRunLoopCtxt pollUntil:within:] (self=0x6989e288, 
_cmd=0x65c623d0, milliseconds=3241, contexts=0x663c3608) at GSRunLoopCtxt.m:633
#25 0x65a54064 in -[NSRunLoop acceptInputForMode:beforeDate:] (self=0x663c3248, 
_cmd=0x65c623f8, mode=0x65c61b18, limit_date=0x66910ba8) at NSRunLoop.m:1198
#26 0x65a54492 in -[NSRunLoop runMode:beforeDate:] (self=0x663c3248, 
_cmd=0x5dd44f78, mode=0x65c61b18, date=0x6a9327a8) at NSRunLoop.m:1266
#27 0x5dc1eaeb in -[WOCoreApplication run] (self=0x65234708, _cmd=0x1c008368) 
at WOCoreApplication.m:576
#28 0x1c002449 in -[SOGo run] (self=0x65234708, _cmd=0x5dd6c420) at SOGo.m:275
#29 0x5dc708f9 in -[WOWatchDog _runChildWithControlSocket:] (self=0x6989b948, 
_cmd=0x5dd6c488, controlSocket=0x6989bc08) at WOWatchDogApplicationMain.m:499
#30 0x5dc70d8e in -[WOWatchDog _spawnChild:] (self=0x6989b948, _cmd=0x5dd6c4f8, 
child=0x6989b2c8) at WOWatchDogApplicationMain.m:563
#31 0x5dc71578 in -[WOWatchDog _ensureChildren] (self=0x6989b948, 
_cmd=0x5dd6c600) at WOWatchDogApplicationMain.m:653
#32 0x5dc729d0 in -[WOWatchDog run:argc:argv:] (self=0x6989b948, 
_cmd=0x5dd6c688,

Re: [SOGo] Weird load issue

2011-11-30 Thread André Schild


Am 29.11.2011 18:29, schrieb Donny Brooks:
For the past few months since deploying thunderbird with the 
connector/lightning/integrator plugins to over half of our users we 
have been having some issues. Mainly, if the load on the server ever 
reaches 1.00 and you try to compose a new email in thunderbird once 
you get the first or second letter in the address bar and it just 
hangs. To exit you have to close the window.

Are you sure it's not this issue ?

https://bugzilla.mozilla.org/show_bug.cgi?id=608041

André
--
users@sogo.nu
https://inverse.ca/sogo/lists

Re: [SOGo] Weird load issue

2011-11-30 Thread Julian Robbins


On 29/11/11 19:02, Donny Brooks wrote:


On Tuesday, November 29, 2011 12:57 PM CST, Georg Bretschneider 
georg+s...@georgb.de wrote:



Hi,


Am 29.11.2011 19:38, schrieb Donny Brooks:
 AH HA!

 [root@webmail ~]# ps aux | grep sogod
 sogo 27641 20.3 0.3 219992 14120 ? S 09:35 35:59
 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
 -WOLogFile /var/log/sogo/sogo.log
 sogo 28260 1.4 1.2 272796 54360 ? Ss 12:20 0:11
 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
 -WOLogFile /var/log/sogo/sogo.log
 root 28323 0.0 0.0 61156 740 pts/1 S+ 12:32 0:00 grep sogod
 [root@webmail ~]#

 I now see that it needed to be changed in /etc/sysconfig/sogo also! 
So I

 changed it to 8 there also and now I see 8 processes.

If I'm not mistaken, this is where you have to change the worker count
when you are starting sogo with the init.d script; according to the
manual, if I remember correctly.


 Could this help solve the issue? I know time will tell but since you
 deal with it day in and day out you may know before anyone Ludovic.
 Also, what is a reasonable number of WOWorkers to use on my setup?
 Virtual machine with 4 cpu cores, 4GB memory and up to 200 people
 hitting it at once.

Not much experience, but I would start with four workers.

 --
 Donny B
 MDAH

 On Tuesday, November 29, 2011 12:25 PM CST, Ludovic Marcotte
 wrote:

 On 29/11/11 13:23, Donny Brooks wrote:
  The only reference to the WOWorkerCount is for the GNUStepDefaults
  file. I am running Centos 5.4 currently on that machine and have no
  /etc/sogo anything.
 And what's the value? What does ps aux | grep sogod tells you?

 Regards,

Bye
Geprg
--
users@sogo.nu
https://inverse.ca/sogo/lists


Well in the past 30 minutes the load has gone from around 0.10-0.20 on 
average to 0.00 and the 15 minute load is at 0.01! I would say this 
resolved the load issue at least. Now just to see if anyone has the 
composition issue.
So it looks like I was on the right lines about defining WOWorkers in 
the /etc config file not with the sogo home users GNUStep file. Its easy 
to miss that with an init script (I know I said daemon but I was rushing 
to leave at the time). The number of WOWorkers is mentioned in the log 
on sogo start up,


ie 0x0x2297ba0[WOWatchDog] preparing 5 children
Nov 30 06:32:41 sogod [28032]: 0x0x2297ba0[WOWatchDog] child spawned 
with pid 28044
Nov 30 06:32:41 sogod [28032]: 0x0x2297ba0[WOWatchDog] child spawned 
with pid 28045
Nov 30 06:32:41 sogod [28032]: 0x0x2297ba0[WOWatchDog] child spawned 
with pid 28046
Nov 30 06:32:41 sogod [28032]: 0x0x2297ba0[WOWatchDog] child spawned 
with pid 28047
Nov 30 06:32:41 sogod [28032]: 0x0x2297ba0[WOWatchDog] child spawned 
with pid 28048


Whereas by default, I think its just set to one WOWorker.

I still have my problem, but perhaps although sounding similar is 
actually something quite different.


Julian



--
Donny B
MDAH



--
Kind Regards,

Julian Robbins
Web Marketing and IT Manager

Q-par Angus Ltd, Barons Cross Laboratories, Leominster, Herefordshire, HR6 8RS, 
United Kingdom

Telephone +44 (0) 1568 612138, Fax +44 (0) 1568 616373, Web www.q-par.com

IDEAS ENGINEERED - SPECIALISTS IN MICROWAVE and RF ENGINEERING

(Registered office)
Incorporated in England No 1826221



--
users@sogo.nu
https://inverse.ca/sogo/lists

[SOGo] Weird load issue

2011-11-29 Thread Donny Brooks

For the past few months since deploying thunderbird with the connector/lightning/integrator plugins to over half of our users we have been having some issues. Mainly, if the load on the server ever reaches 1.00 and you try to compose a new email in thunderbird once you get the first or second letter in the address bar and it just hangs. To exit you have to close the window.

When the load reaches that high it is always a single sogod process eating up 98%-100% of one of the cpu cores. The machine is a Xen paravirtualized machine with 4 cpu cores and 4GB ram allocated to it. I have to stop and restart httpd/sogod/memcached to get the process to behave.

The high load affects all users no matter if they access via the web interface or Thunderbird. The only ones not affected are the users that use Thunderbird without the plugins. I have set the WOWorkersCount to 8 for testing but that does not seem to have done anything either way.

Does anyone have any idea of what could be going on and/or what to do? Since we are rolling this out to all users in one form or another (web interface or thunderbird) now this is imperative I find out whats going on soon. The logs dont seem to be any help at this point but I am still digging through them. Thanks in advace for any help on this.
--
Donny B
MDAH

Re: [SOGo] Weird load issue

2011-11-29 Thread Julian Robbins


On 29/11/11 17:29, Donny Brooks wrote:
For the past few months since deploying thunderbird with the 
connector/lightning/integrator plugins to over half of our users we 
have been having some issues. Mainly, if the load on the server ever 
reaches 1.00 and you try to compose a new email in thunderbird once 
you get the first or second letter in the address bar and it just 
hangs. To exit you have to close the window.


We've seen exactly this happen a few times, with same setup, ie sogo 
with full connector/lightning/integrator plugins. But I don't think its 
related to load (for us). We only have 35 users so the sogo server load 
is usually pretty low.


We've found that Windows 7 users haven't reported this issue as yet, 
wheras with XP its happened a few times. (We have no Vista users). But 
this may be related to the specs of the machines being better with 7.  
That said we have some uers with slow older Win XP machines that havent 
seen this issue.


It appears that Thunderbird 3 is poor with respect to handling Address 
Books via Card-dav. Its been mentioned on the list a few times before. 
Thunderbird 2 was ok apparently.


Have you tried using the Sogo Integrator autocomplete delay adjustment  
user_pref(sogo-connector.autoComplete.delay, 1300); as mentioned in a 
bug report (sorry not got time to search for it). It helps stopping 
double letters appearing but may be a bit high in regard to timeouts ... 
We have it included in the Integrator config so its rolled out into 
everyone's prefs.js thunderbird file.




When the load reaches that high it is always a single sogod process 
eating up 98%-100% of one of the cpu cores. The machine is a Xen 
paravirtualized machine with 4 cpu cores and 4GB ram allocated to it. 
I have to stop and restart httpd/sogod/memcached to get the process to 
behave.
Are you really using WO workers count correctly? You have to set it 
somewhere in /etc/sogo not in the normal user's GNUStep conf file if 
running as a daemon. It's in the manual, but easy to miss.


The high load affects all users no matter if they access via the web 
interface or Thunderbird. The only ones not affected are the users 
that use Thunderbird without the plugins. I have set the 
WOWorkersCount to 8 for testing but that does not seem to have done 
anything either way.
Generally web users aren't affected here, it always seems to work fine 
for us.


Does anyone have any idea of what could be going on and/or what to do? 
Since we are rolling this out to all users in one form or another (web 
interface or thunderbird) now this is imperative I find out what's 
going on soon. The logs don't seem to be any help at this point but I 
am still digging through them. Thanks in advace for any help on this.
I couldnt see anything in the logs either on this either. Please let me 
know if you get anywhere with this, if you find out anything offlist as 
I'd like to know.



Julian Robbins
--
users@sogo.nu
https://inverse.ca/sogo/lists

Re: [SOGo] Weird load issue

2011-11-29 Thread Donny Brooks

On Tuesday, November 29, 2011 11:42 AM CST, Julian Robbins julian.robb...@q-par.com wrote:

On 29/11/11 17:29, Donny Brooks wrote:
For the past few months since deploying thunderbird with the
connector/lightning/integrator plugins to over half of our users we
have been having some issues. Mainly, if the load on the server ever
reaches 1.00 and you try to compose a new email in thunderbird once
you get the first or second letter in the address bar and it just
hangs. To exit you have to close the window.

Weve seen exactly this happen a few times, with same setup, ie sogo
with full connector/lightning/integrator plugins. But I dont think its
related to load (for us). We only have 35 users so the sogo server load
is usually pretty low.

Weve found that Windows 7 users havent reported this issue as yet,
wheras with XP its happened a few times. (We have no Vista users). But
this may be related to the specs of the machines being better with 7.
That said we have some uers with slow older Win XP machines that havent
seen this issue.

It appears that Thunderbird 3 is poor with respect to handling Address
Books via Card-dav. Its been mentioned on the list a few times before.
Thunderbird 2 was ok apparently.

Have you tried using the Sogo Integrator autocomplete delay adjustment
user_pref(sogo-connector.autoComplete.delay, 1300); as mentioned in a
bug report (sorry not got time to search for it). It helps stopping
double letters appearing but may be a bit high in regard to timeouts ...
We have it included in the Integrator config so its rolled out into
everyones prefs.js thunderbird file.

When the load reaches that high it is always a single sogod process
eating up 98%-100% of one of the cpu cores. The machine is a Xen
paravirtualized machine with 4 cpu cores and 4GB ram allocated to it.
I have to stop and restart httpd/sogod/memcached to get the process to
behave.
Are you really using WO workers count correctly? You have to set it
somewhere in /etc/sogo not in the normal users GNUStep conf file if
running as a daemon. Its in the manual, but easy to miss.

The high load affects all users no matter if they access via the web
interface or Thunderbird. The only ones not affected are the users
that use Thunderbird without the plugins. I have set the
WOWorkersCount to 8 for testing but that does not seem to have done
anything either way.
Generally web users arent affected here, it always seems to work fine
for us.

Does anyone have any idea of what could be going on and/or what to do?
Since we are rolling this out to all users in one form or another (web
interface or thunderbird) now this is imperative I find out whats
going on soon. The logs dont seem to be any help at this point but I
am still digging through them. Thanks in advace for any help on this.
I couldnt see anything in the logs either on this either. Please let me
know if you get anywhere with this, if you find out anything offlist as
Id like to know.

Julian Robbins
--
users@sogo.nu
https://inverse.ca/sogo/lists
The load normally is around 0.10 to 0.20 at any given point on our server. It is only once a rogue sogod process starts taking up the available cpu cycles that this happens. I have applied a fix for the timeout via a thunderbird.cfg file that we force preferences to the end users. We will see how that works out.

As for the machines being affected, all of ours is Vista or 7. Actually all of the ones now are 7 since we just deployed a whole slew of new pcs and that is when we setup the plugins too. So I dont think it is end user hardware that is totally at fault.

The only reference to the WOWorkerCount is for the GNUStepDefaults file. I am running Centos 5.4 currently on that machine and have no /etc/sogo anything.

The web users problem is not just the address book lookup but the whole web interface freezes. You cannot load new messages or change folders when this happens.

So with the change on the timeout I am going to see how this works. Also I have a small test group that I am disabling the plugins just to make sure it isnt thunderbird also.
--
Donny B
MDAH

Re: [SOGo] Weird load issue

2011-11-29 Thread Ludovic Marcotte


On 29/11/11 13:23, Donny Brooks wrote:
The only reference to the WOWorkerCount is for the GNUStepDefaults 
file. I am running Centos 5.4 currently on that machine and have no 
/etc/sogo anything.

And what's the value? What does ps aux  | grep sogod tells you?

Regards,

--
Ludovic Marcotte
lmarco...@inverse.ca  ::  +1.514.755.3630  ::  www.inverse.ca
Inverse inc. :: Leaders behind SOGo (www.sogo.nu) and PacketFence 
(www.packetfence.org)

--
users@sogo.nu
https://inverse.ca/sogo/lists

Re: [SOGo] Weird load issue

2011-11-29 Thread Donny Brooks

AH HA!

[root@webmail ~]# ps aux | grep sogod
sogo 27641 20.3 0.3 219992 14120 ? S 09:35 35:59 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid -WOLogFile /var/log/sogo/sogo.log
sogo 28260 1.4 1.2 272796 54360 ? Ss 12:20 0:11 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid -WOLogFile /var/log/sogo/sogo.log
root 28323 0.0 0.0 61156 740 pts/1 S+ 12:32 0:00 grep sogod
[root@webmail ~]#

I now see that it needed to be changed in /etc/sysconfig/sogo also! So I changed it to 8 there also and now I see 8 processes.

Could this help solve the issue? I know time will tell but since you deal with it day in and day out you may know before anyone Ludovic. Also, what is a reasonable number of WOWorkers to use on my setup? Virtual machine with 4 cpu cores, 4GB memory and up to 200 people hitting it at once.
--
Donny B
MDAH

On Tuesday, November 29, 2011 12:25 PM CST, Ludovic Marcotte lmarco...@inverse.ca wrote:


	On 29/11/11 13:23, Donny Brooks wrote:
	 The only reference to the WOWorkerCount is for the GNUStepDefaults
	 file. I am running Centos 5.4 currently on that machine and have no
	 /etc/sogo anything.
	And whats the value? What does ps aux | grep sogod tells you?
	
	Regards,
	
	--
	Ludovic Marcotte
	lmarco...@inverse.ca :: +1.514.755.3630 :: www.inverse.ca
	Inverse inc. :: Leaders behind SOGo (www.sogo.nu) and PacketFence (www.packetfence.org)
	
	--
	users@sogo.nu
	https://inverse.ca/sogo/lists

Re: [SOGo] Weird load issue

2011-11-29 Thread Georg Bretschneider

Hi,


Am 29.11.2011 19:38, schrieb Donny Brooks:
 AH HA!
 
 [root@webmail ~]# ps aux | grep sogod
 sogo 27641 20.3  0.3 219992 14120 ?S09:35  35:59
 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
 -WOLogFile /var/log/sogo/sogo.log
 sogo 28260  1.4  1.2 272796 54360 ?Ss   12:20   0:11
 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
 -WOLogFile /var/log/sogo/sogo.log
 root 28323  0.0  0.0  61156   740 pts/1S+   12:32   0:00 grep sogod
 [root@webmail ~]#
 
 I now see that it needed to be changed in /etc/sysconfig/sogo also! So I
 changed it to 8 there also and now I see 8 processes.

If I'm not mistaken, this is where you have to change the worker count
when you are starting sogo with the init.d script; according to the
manual, if I remember correctly.

 
 Could this help solve the issue? I know time will tell but since you
 deal with it day in and day out you may know before anyone Ludovic.
 Also, what is a reasonable number of WOWorkers to use on my setup?
 Virtual machine with 4 cpu cores, 4GB memory and up to 200 people
 hitting it at once.

Not much experience, but I would start with four workers.

 --
 Donny B
 MDAH
 
 On Tuesday, November 29, 2011 12:25 PM CST, Ludovic Marcotte
 lmarco...@inverse.ca wrote:
 
 On 29/11/11 13:23, Donny Brooks wrote:
  The only reference to the WOWorkerCount is for the GNUStepDefaults
  file. I am running Centos 5.4 currently on that machine and have no
  /etc/sogo anything.
 And what's the value? What does ps aux | grep sogod tells you?

 Regards,

Bye
Geprg
-- 
users@sogo.nu
https://inverse.ca/sogo/lists

Re: [SOGo] Weird load issue

2011-11-29 Thread Donny Brooks


On Tuesday, November 29, 2011 12:57 PM CST, Georg Bretschneider georg+s...@georgb.de wrote:


	Hi,
	
	
	Am 29.11.2011 19:38, schrieb Donny Brooks:
	 AH HA!
	
	 [root@webmail ~]# ps aux | grep sogod
	 sogo 27641 20.3 0.3 219992 14120 ? S 09:35 35:59
	 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
	 -WOLogFile /var/log/sogo/sogo.log
	 sogo 28260 1.4 1.2 272796 54360 ? Ss 12:20 0:11
	 /usr/sbin/sogod -WOWorkersCount 1 -WOPidFile /var/run/sogo/sogo.pid
	 -WOLogFile /var/log/sogo/sogo.log
	 root 28323 0.0 0.0 61156 740 pts/1 S+ 12:32 0:00 grep sogod
	 [root@webmail ~]#
	
	 I now see that it needed to be changed in /etc/sysconfig/sogo also! So I
	 changed it to 8 there also and now I see 8 processes.
	
	If Im not mistaken, this is where you have to change the worker count
	when you are starting sogo with the init.d script; according to the
	manual, if I remember correctly.
	
	
	 Could this help solve the issue? I know time will tell but since you
	 deal with it day in and day out you may know before anyone Ludovic.
	 Also, what is a reasonable number of WOWorkers to use on my setup?
	 Virtual machine with 4 cpu cores, 4GB memory and up to 200 people
	 hitting it at once.
	
	Not much experience, but I would start with four workers.
	
	 --
	 Donny B
	 MDAH
	
	 On Tuesday, November 29, 2011 12:25 PM CST, Ludovic Marcotte
	  wrote:
	
	 On 29/11/11 13:23, Donny Brooks wrote:
	  The only reference to the WOWorkerCount is for the GNUStepDefaults
	  file. I am running Centos 5.4 currently on that machine and have no
	  /etc/sogo anything.
	 And whats the value? What does ps aux | grep sogod tells you?
	
	 Regards,
	
	Bye
	Geprg
	--
	users@sogo.nu
	https://inverse.ca/sogo/lists

Well in the past 30 minutes the load has gone from around 0.10-0.20 on average to 0.00 and the 15 minute load is at 0.01! I would say this resolved the load issue at least. Now just to see if anyone has the composition issue.

--
Donny B
MDAH

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

[SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

Re: [SOGo] Weird load issue

10 matches

Site Navigation

Mail list logo

Footer information