'Re: [ngw] GW 8.0.3 Horrible Slowness'

[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ngw
Subject:    Re: [ngw] GW 8.0.3 Horrible Slowness
From:       "Hamish Speirs" <hamish () haitch ! net>
Date:       2012-11-28 0:51:56
Message-ID: 50B50BCC0200005600054595 () postie ! caledonia ! net
[Download RAW message or body]

The two most common issues I run into on VMWare setups are:

1) VMWare has silently set a memory resource limit on the VM. This will cripple the \
VM under any sort of moderate load. To check the memory resource settings on the VM - \
go into Edit Settings for the VM, then go to the resources tab, then the memory \
resources setting. Make sure the memory limit is set to either unlimited, or at least \
the same amount of memory as is assigned to the VM. If it's been limited, set it back \
to unlimited, then reboot the VM asap.

2) The iSCSI networking setup is really badly done. Properly configured iSCSI can be \
extremely fast, improperly configured iSCSI is horrible.

Renee - What sort of SAN? How many NICs for iSCSI on the hosts and on the SAN? \
Software iSCSI initiator or hardware iSCSI HBAs?

Hamish.

> > > "Renee Hunter" <Renee@wfse.org> 11/27/2012 2:52 PM >>>
Our drives in the hosts and SAN are SAS.  What do you recommend?

> > > "Elbert LaGrew" <Elbert.LaGrew@katun.com> 11/27/2012 7:10 AM >>>
I'd stay away from SAS like the plague...had nothing but grief with GW on those \
things.....

--El

> > > On 11/27/2012 at 9:07 AM, in message \
> > > <50B490FD020000750003E7E0@inet.eastcobbgroup.com>, "James Taylor" \
> > > <James.Taylor@eastcobbgroup.com> wrote:

I haven't seen any reason to use fibrechannel in any of our new installations.
It is expensive, not very flexible and I don't believe it has any performance \
advantage. We continue to recommend iSCSI, barring any specific considerations.
-jt

> > > "Renee Hunter" <Renee@wfse.org> 11/27/2012   01:11 AM >>> 
What do you use now?  Novell Sr Tech today recommended Fiber Channel.

> > > "James Taylor" <James.Taylor@eastcobbgroup.com> 11/26/2012 8:40 PM >>>
The system I had the problem with was iSCSI attached, but after wasting a lot of time \
checking the switch, frame settings and network drivers, it ended up being the \
physical disk speed. iSCSI should not be a bottleneck if the LAN is spec'ed and \
                configured properly.
-jt

> > > "Renee Hunter" <Renee@wfse.org> 11/26/2012   10:37 PM >>> 
Wondering what is used in virtual environments.  Ours was set up with iSCSI to the \
SANs.

> > > "James Taylor" <James.Taylor@eastcobbgroup.com> 11/26/2012 6:31 PM >>>
It's good you have that option.
If it is disk io, then you should see some immediate relief.
-jt 

> > > "Renee Hunter" <Renee@wfse.org> 11/26/2012   08:26 PM >>> 
The VM Host is SAS 10K. Will try later tonight to migrate GW to a faster host - 15K.
HP Storage Works SAN drives are SAS 15K
125 users 1 PO

> > > "James Taylor" <James.Taylor@eastcobbgroup.com> 11/26/2012 5:08 PM >>>
What is your disk drive configuration?
Specifically the type and speed of the disk drives?
Also, how many users?
I had a similar occurrence a few years ago. Every thing had been working fine for \
years, and seemingly overnight, the GroupWise performance hit a wall. After weeks of \
trying various configuration changes, we found out that the GW database I/O had \
exceeded the max amount of data that could be physically transferred from the 10K rpm \
SCSI drives we were using. The only resolution for the problem was to replace the \
                drives with 15k rpm equivalents.
-jt

James Taylor
678-697-9420
james.taylor@eastcobbgroup.com 

> > > "Renee Hunter" <Renee@wfse.org> 11/26/2012   06:53 PM >>> 
GW 8.0.3 Clients 8.0.2 Vmware 4.1, SUSE Linux 10, OES 2.0.3 Patch Level 3  

GW and the gw server performance are fine until around 10-10:30am when users start \
hitting GW.  Then it and the server become incredibly slow. GW client stops \
responding, locks up, etc.  Opening a terminal window is substantially slower and \
running a status shows all parts are running, but is incredible slow responding after \
10am.  Neither restarting GW nor the server help.  Performance improves as uses log \
off for the night.

Have been working with Novell tech spt for two weeks with no resolution yet.  \
Escalated earlier today but was told not to expect a call back until probably \
tomorrow.  Still hoping for today. 

Would appreciate any ideas of what might be affecting us so much!

If this is an unsolicited spam message, please click this link to report it: \
http://control.eastcobbgroup.com:49285/contents/spamreport.shtml?rptid=17862&srvid=16vl15t \

_______________________________________________
ngw mailing list
ngw@ngwlist.com 
http://ngwlist.com/mailman/listinfo/ngw 

If this is an unsolicited spam message, please click this link to report it: \
http://control.eastcobbgroup.com:49285/contents/spamreport.shtml?rptid=17870&srvid=16vl15t \

_______________________________________________
ngw mailing list
ngw@ngwlist.com 
http://ngwlist.com/mailman/listinfo/ngw 

_______________________________________________
ngw mailing list
ngw@ngwlist.com 
http://ngwlist.com/mailman/listinfo/ngw 

_______________________________________________
ngw mailing list
ngw@ngwlist.com 
http://ngwlist.com/mailman/listinfo/ngw 
_______________________________________________
ngw mailing list
ngw@ngwlist.com 
http://ngwlist.com/mailman/listinfo/ngw
_______________________________________________
ngw mailing list
ngw@ngwlist.com
http://ngwlist.com/mailman/listinfo/ngw

[prev in list] [next in list] [prev in thread] [next in thread] 