OpenVZ Forum


Home » General » Support » Zimbra on OpenVz
Zimbra on OpenVz [message #38345] Wed, 09 December 2009 13:10 Go to next message
burn is currently offline  burn
Messages: 4
Registered: December 2009
Junior Member
I tried to install Zimbra 6.0.3 into OpenVZ container, and failed. Google search led me to http://wiki.openvz.org/Zimbra_on_OpenVZ_on_CentOS , which suggests increasing memory pages until failcnt drops to zero. It did the trick, but if I understand correctly, increasing memory beyond available leads to potentially unstable configuration. And in the wiki example there is about 8Tb of privvmpages granted to the zimbra container. Could anyone clarify? Is it a recommended setup? Any personal experience with running Zimbra in OpenVZ would be welcome, too.
Re: Zimbra on OpenVz [message #38347 is a reply to message #38345] Wed, 09 December 2009 13:22 Go to previous messageGo to next message
kir is currently offline  kir
Messages: 1645
Registered: August 2005
Location: Moscow, Russia
Senior Member

That is why I have added a warning to that page:

Quote:
Warning: The above UBC values might not be consistent; see UBC consistency check.


You have to learn UBCs in order to be able to set them up correctly. But for the test run on a test server you can use unlimited configuration.


Kir Kolyshkin
http://static.openvz.org/userbars/openvz-developer.png
Re: Zimbra on OpenVz [message #38349 is a reply to message #38347] Wed, 09 December 2009 13:57 Go to previous messageGo to next message
burn is currently offline  burn
Messages: 4
Registered: December 2009
Junior Member
I've read it, and by these criterias 8 Tb of privvmpages is pretty consistent. Still, I have doubts that such extreme overcommitment of allocated memory is stable.
Re: Zimbra on OpenVz [message #38418 is a reply to message #38349] Fri, 18 December 2009 10:16 Go to previous messageGo to next message
batfastad is currently offline  batfastad
Messages: 15
Registered: September 2009
Location: London, UK
Junior Member
Hi burn

Have you actually been able to get Zimbra to run without failcnt on privvmpages?

I'm trying to do the same thing... but no matter what value of privvmpages I enter as barrier & limit, I still get a failcnt of about ~17600 for privvmpages. Even if I enter it as ridiculously high per that example in the wiki.

Cheers, B
Re: Zimbra on OpenVz [message #38421 is a reply to message #38418] Fri, 18 December 2009 12:41 Go to previous messageGo to next message
burn is currently offline  burn
Messages: 4
Registered: December 2009
Junior Member
I haven't yet put it into production, but it runs for a week now, and failcnt is still 0. I'm using the latest Proxmox VE. The container is Centos 5.4 x86_64, Zimbra version 6.0.3 OSE. Both ram and swap are set to 2,5 Gb in Proxmox interface, and I manually increased privvmpages up to 64 Gb.
Re: Zimbra on OpenVz [message #38424 is a reply to message #38345] Fri, 18 December 2009 12:57 Go to previous messageGo to next message
batfastad is currently offline  batfastad
Messages: 15
Registered: September 2009
Location: London, UK
Junior Member
Hi burn
Thanks for the reply. So when you say you upped privvmpages to 64Gb, what was the actual value you used?
I'm still getting confused by the maths in converting MB/GB to pages

I'm not running Proxmox, just plain CentOS 5.4 and OpenVZ... kind of wish I was running Proxmox now though Sad

[Updated on: Fri, 18 December 2009 12:57]

Report message to a moderator

Re: Zimbra on OpenVz [message #38429 is a reply to message #38424] Fri, 18 December 2009 14:00 Go to previous message
burn is currently offline  burn
Messages: 4
Registered: December 2009
Junior Member
One privvmpage unit is 4 kb of memory. Here's the entire container config, see if it helps
amber:/etc/vz/conf# cat 107215.conf
# PVE default config for 256MB RAM

ONBOOT="no"

# Primary parameters
NUMPROC="1024:1024"
NUMTCPSOCK="9223372036854775807:9223372036854775807"
NUMOTHERSOCK="9223372036854775807:9223372036854775807"
VMGUARPAGES="1310720:9223372036854775807"

# Secondary parameters
KMEMSIZE="9223372036854775807:9223372036854775807"
OOMGUARPAGES="1310720:9223372036854775807"
PRIVVMPAGES="15728640:16777216"
TCPSNDBUF="9223372036854775807:9223372036854775807"
TCPRCVBUF="9223372036854775807:9223372036854775807"
OTHERSOCKBUF="9223372036854775807:9223372036854775807"
DGRAMRCVBUF="9223372036854775807:9223372036854775807"

# Auxiliary parameters
NUMFILE="9223372036854775807:9223372036854775807"
NUMFLOCK="9223372036854775807:9223372036854775807"
NUMPTY="255:255"
NUMSIGINFO="1024:1024"
DCACHESIZE="9223372036854775807:9223372036854775807"
LOCKEDPAGES="655360:655360"
SHMPAGES="9223372036854775807:9223372036854775807"
NUMIPTENT="9223372036854775807:9223372036854775807"
PHYSPAGES="0:9223372036854775807"

# Disk quota parameters
DISKSPACE="67108864:73819750"
DISKINODES="12800000:14080000"
QUOTATIME="0"
QUOTAUGIDLIMIT="0"

# CPU fair sheduler parameter
CPUUNITS="1000"
CPUS="1"
VE_ROOT="/var/lib/vz/root/$VEID"
VE_PRIVATE="/var/lib/vz/private/$VEID"
OSTEMPLATE="centos-5-standard_5.4_amd64"
ORIGIN_SAMPLE="pve.auto"
IP_ADDRESS="192.168.107.215"
HOSTNAME="mail.domain.com"
NAMESERVER="192.168.102.5 192.168.101.112"
SEARCHDOMAIN="domain.com"
Previous Topic: netdev_add kills wanpipe ??
Next Topic: High Virtual Memory Usage when loading shared libraries
Goto Forum:
  


Current Time: Tue Dec 03 08:49:12 GMT 2024

Total time taken to generate the page: 0.11388 seconds