VMware Cloud Community
noesberger
Contributor
Contributor

Use Update Manager over a slow WAN Link

We have the Virtual Center installed in Switzerland and are managing ESX Server in whole Europe.

When I try to use the Update Manager to Patch the ESX-Server on a Remote Location (Connected over a 2-3 Mbit's Network) I receive always the following error in the Virtual Center:

Task: Install, Status: Operation timed out.

When I check the log on the ESX Server directly (esxupdate.log) I can see that it's still downloading the file from the Update Manager.

Can I confiure somewhere the time, that the Update-Manager/esxupdate/VC wait's until the operation time outs.

Because always after 15 Minutes I receive the error, because the patch is not yet copied to the esx-server.

Reply
0 Kudos
14 Replies
petkom
Community Manager
Community Manager

Hi,

if you are with the latest VUM and vSphere then your problem is gone! You can use the stage functionality that is moving the patches for remediation from VUM repository to the hosts. When the patches are staged you can do the remediation with minimal downtime.

If you don't have the latest version check vci-integrity.xml file for various timeout options but I'm not sure if this exactly is present there. It is in C:\Program Files\VMware\Infrastructure\Update Manager\

hope this helps

thanks,

Petko

noesberger
Contributor
Contributor

Thanks for your fast reply.

It's not planed to migrate to vSphere in the next 12 months. So it would be good to find a solution for my problem. But a question to vSphere I have. Can you select the directory to stage the patches. Because to store all patches the space on the "default" partitions could be too small, or can you choose a VMFS Datastore to save the file temporarily.

I found a section in the vci-integrity.xml file, that could be very intersting to solve the problem:

<valm>

<disableValm>false</disableValm>

<encryptionKey>ssl/valm.key</encryptionKey>

<getUpdateTimeout>300</getUpdateTimeout>

<pollInterval>5</pollInterval>

<remediateTimeout>1200</remediateTimeout>

</valm>

Does someone know what VALM is for?

Thanks for your help.

Reply
0 Kudos
petkom
Community Manager
Community Manager

Hi again,

In VMware terms VALM usually stands for VMware Appliance Lifecycle Manager - it's about virtual appliances

About the stages in vSphere. I think you cannot change the default path. The staged patches are stored under "/tmp/updatecache" on the host. All the staged patches will be cleared after a remediation operation. So the data will not accumulate as long as users do staging and remediation intermittently. The default partitioning should suffice. (I think that /tmp/updatecache is mounted on a partition which has about 4GB free space, you should probably check ESX installation guide for the exact number)

Also, VUM has a safety check, it will check ESX/ESXi for avaiable space for "/tmp/updatecache" before staging so that space utilization on the host is not exceeded. If there's not enough physical disk space, vUM will request ESX to allocatea a ramdisk for "/tmp/updatecache". If even that fails (due to insufficient RAM) VUM will report error back to user.

hph

thanks,

Petko

Reply
0 Kudos
noesberger
Contributor
Contributor

Hi,

I tried to change the parameters, but the error is the same. Exactly after 15 mintues it cancels the task.

<getUpdateTimeout>*300*</getUpdateTimeout>

<remediateTimeout>*1200*</remediateTimeout>

Reply
0 Kudos
petkom
Community Manager
Community Manager

Hi noesberger,

I've checked your issue and the answer is:

  • you can set timeout values in vci-integrity.xml for patch download but only in VUM 4.0 and for 4.0 hosts

one suggestion - try to put the host manually in maintanance mode and then start remediation - the performance should be better

thanks,

Petko

Reply
0 Kudos
JHewson
Contributor
Contributor

HI, what timeouts are you actually changing? I have a low wan link in the test lab and find when the host has to have updates over this it keeps timing out ofter 15 minutes, regardless if I am remediating or staging. I do have 4.0 hosts in the test area and its not going to be practical to put an update server out at the branch, so would like to get it working across the WAN link. This even happens on small update, only a few hundred kb too.

Reply
0 Kudos
noesberger
Contributor
Contributor

VMWare has created a KB Article for this problem. I can now install the ESX Patches also over slow WAN links.

(Problem solved!!!)

Reply
0 Kudos
steveanderson3
Contributor
Contributor

We are experiencing this issue, but we are using ESXi 4 hosts, not ESX. The KB article mentions making a modification on the esx host as well. I'm going to test this without the ESX host-side mod.

Reply
0 Kudos
vSeanClark
Enthusiast
Enthusiast

petfom,

Staging is nice but have you used it over a slow pipe? Even staging was timing out for us last night. Trying to stage the July ESX 4 updates over a T1 would timeout. still looking for my best solution here. Might be esxupdate. Smiley Sad

If you value your karma, please consider awarding points for "Correct" or "Helpful". Smiley Happy

Sean Clark - vExpert, VCP - - <!Session data>

UPDATE: going to try this tomorrow to see if it fixes our issue: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100967...

<!Session data>

Sean Clark - http://twitter.com/vseanclark
Reply
0 Kudos
ssauer
Contributor
Contributor

Hitting same issue's, some of our remote sites are no problem but those that have higher latency are timing out. I am trying the proposed fix to see if it helps.

Reply
0 Kudos
ssauer
Contributor
Contributor

Fixed our issue with the posted solution (vSphere 4).

Reply
0 Kudos
MartinHolt
Contributor
Contributor

Hi, I am experiencing this exact problem at a Client Site. They are running the latest version of vSphere.

I get a message saying "you are not authorised to access this document" error when I try and look at the proposed solution KB article (). Can anyone suggest either why this is happenning or give a better idea of what the solution is so I can try this with my client.

Cheers

Martin

Reply
0 Kudos
roity57
Enthusiast
Enthusiast

I tried this fix and it did not work. It persists on timing out after 15 minutes. I edited the vpxd.cfg as specified and restarted Virtual Center server, and tried this twice and even had to restart the Update Manager at one point.

Can someone please detail the fix as they implemented it so I can see whether or not I've mis understood VMWares KB article?

So far I can remediate hosts, I have Update 4 but 5 security patches pending. If I remediate about 3 times it finally finishes and I too have had to manually kill the process running on the ESX host in the past but if I let it time out and wait a while it finishes the patch install anyway (a bit like committing big snapshots and timing out but finishing the commit) so eventually I can remediate the host completely.

Failing that it's off to VMWare support Smiley Happy

Thanks,

Dave

Reply
0 Kudos