Problems with Outlook after upgrade to 6.0.6

[Labsy]

Hi,

I am not sure what to relate my problems with, but here is how it looks:
- I have about 30 domains on ZCS 6.0.6, and about 300 users total
- some of them have over 8 GB mails in their mailbox
- most of them use OUTLOOK versions 2003 to 2010
- most of them use IMAP+SSL to access their mailbox

But only some of them, regardless of how full their mailbox is, have random problems with using Outlook.
It WAS WORKING FINE for about half a year, but last 2-3 weeks many of users report very really VERY slow response of Outlook, for example:
- when user clicks on another message, it takes from few seconds to FEW MINUTES for Outlook to show new message
- some of them have noticed Outlook saying "Cannot upload headers", but they are not sure of exact error
- I personally use WEBMAIL all the time, and I have noticed that changing folders in Webmail is something like slowish; it takes 2-5 seconds for Webmail to switch between inbox folders, and I have about 2 GB of mails there

I'd need an advice where to start looking?
What should I check?

Server is ESX, but has 32 GB of RAM and RAID 10 disk array, and QuadXEON CPU is about 8% average busy, so I think hardware should not be a problem.

Any advice appreciated.
Thanx!

[phoenix]

Server is ESX, but has 32 GB of RAM and RAID 10 disk array, and QuadXEON CPU is about 8% average busy, so I think hardware should not be a problem.

Yes but you haven't given any details of what the VM is that Zimbra is running on nor what resources the VM has. You also haven't mentione what other VMs are running on the ESX host and what the load is on those VMs.

You could also look at top for starters to see what's happening on the server when the response is 'poor' for the Outlook users. You could also look at the performance monitors in ESX and see what's happening on the server. Does it get worse the longer the server is up or is it always poor? You could always look at your network configuration for bottlenecks, you could have a look in the wiki at the article on Zimbra performance inside a VMware Virtual Machine.

[Labsy]

Phoenix, you are blazing fast with your replies! Thanx.

No, I did not get into details about CPU, HardDisks and RAM on ESX server, because from my sight of view it looks pretty fine:

On Zimbra itself, running without visible problems since last update from 6.0.3 to 6.0.6:

top - 15:39:33 up 35 days,  7:56,  1 user,  load average: 0.38, 0.48, 0.52
Tasks: 141 total,   1 running, 140 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.2%us,  0.2%sy,  0.0%ni, 99.5%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   3872628k total,  3597688k used,   274940k free,   141860k buffers
Swap:  9289720k total,  1497788k used,  7791932k free,   633676k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 8087 zimbra    20   0 45340 4240 2100 S    0  0.1  12:08.50 zmstat-proc
30224 labsy     20   0 18992 1280  932 R    0  0.0   0:00.01 top
    1 root      20   0  4020  616  536 S    0  0.0   0:02.39 init
    2 root      15  -5     0    0    0 S    0  0.0   0:00.00 kthreadd
    3 root      RT  -5     0    0    0 S    0  0.0   1:30.45 migration/0
    4 root      15  -5     0    0    0 S    0  0.0   0:09.94 ksoftirqd/0
    5 root      RT  -5     0    0    0 S    0  0.0   0:00.20 watchdog/0
    6 root      RT  -5     0    0    0 S    0  0.0   1:28.39 migration/1
    7 root      15  -5     0    0    0 S    0  0.0   0:06.70 ksoftirqd/1
    8 root      RT  -5     0    0    0 S    0  0.0   0:00.29 watchdog/1
    9 root      RT  -5     0    0    0 S    0  0.0   1:26.82 migration/2
   10 root      15  -5     0    0    0 S    0  0.0   0:23.93 ksoftirqd/2
   11 root      RT  -5     0    0    0 S    0  0.0   0:00.17 watchdog/2
   12 root      RT  -5     0    0    0 S    0  0.0   1:32.33 migration/3
   13 root      15  -5     0    0    0 S    0  0.0   0:06.59 ksoftirqd/3
   14 root      RT  -5     0    0    0 S    0  0.0   0:00.17 watchdog/3
   15 root      15  -5     0    0    0 S    0  0.0   0:26.18 events/0

On ESX looking all 3 machines and among all 4 processors:

ESX CPU usage in percent
CPU  Latest  Max    Min    Average
0      5,89    66,47  4,48   20,474
1      6,24    65,48  5,02   21,053
2      5,55    66,5   4,77   20,529
3      5,73    65,26  4,89   20,915

ZIMBRA Virtual Machine CPU usage in Percent:
Latest  Max    Min    Average
0,5     43,63  0,36  11,732

ZIMBRA has 4 GB RAM allocated, but actively uses only 532 MB.
Disk Read or Write does not exceed 50 kB/s, and is mostly idling.


Since nothing special is going on on ESX, and also problems occure ONLY RANDOM and only to some rare users, I think there is something on relation between OUTLOOK --> IMAP --> ZIMBRA.
Maybe Outlook does not handle large mailboxes fine?
Or there is some performance impact in Outlook due to "Check for new mail every XY minutes" settings?

I'll try Googling some more...

[Labsy]

Hmmm....just now I figured out one very specific detail, which might be the cause of the problems:

As I said, problems occure only on specific computer in 2 Companies. Both are my customers, but beside that, Company ONE has also a service contract with Company TWO.
Ok, that's fine, but what do contrats have to do with my Outlook-Zimbra problem?

Both companies have their own Windows DOMAIN servers, and technicians and engeneers from comapny TWO often connect via VPN to Company ONE, so their tech guys can repair problems remotelly.
And very often they FORGET TO DISCONNECT VPN tunnel!

My guess is that it is possible that when VPN is established from Company ONE, let's say XP computer to Company TWO Windows Domanin SERVER RRAS....hmmm, yes, it is possible that some or even ALL the traffic goes through VPN tunnel.
Worst case scenario is that engineers from Company TWO have Server 2008 for their workstation, and in this case VPN is established between 2 Windows servers, so 2008 server is "stronger", both have DNS and NetBIOS enabled, so the stronger one takes precedence.

Might it be that all mail goes through VPN tunnel? And all DNS traffic, even all HTTP traffic?
Despite of different local gateways?

As of now, only this VPN relation is found as common to both customers, having both problems with Zimbra-Outlook ombination.