[robemorr]

Has anyone else seen a problem with a virtual BES stop responding for no apparent reason when it's running on VMware ESX server 3.5 build 110268? It seems some of the Blackberry logs just stop even though the service is still running. Also, at this point, you cannot remotely logon to the VM using TS or the Virtual Center console. This has happened on 4 different Virtual BES and there doesn't appear to be a pattern to it. We have also moved the VMs to a different ESX host and the problem still persists. Any insight and/or similiar experience is appreciated.

P.S. BES version is 4.1.6, user count ~ 150, remote SQL server cluster, everything on local LAN.

[gibson_hg]

If you cannot login to the VM sounds more like an issue with the VM than the BES itself. Does anything else stop working on the VM other than BES?

[robemorr]

Nothing else appears to stop...not that much of anything else is running except the standard Windows services. And when I say you can't logon to the VM...it is that the remote session is unresponsive (i.e. screen is blank and won't accept mouse/kbd input).

I tend to agree with you that it sounds like a VMware problem but the problem only manifest itself when the VM is running BES...none of our other VM servers do this (Virtual DC's etc.)...and they are all pretty much the same except for the BES components.

[jibi]

what kind of specs do you have on said VM? i've had this happen on other types of VMs and i have to shut it down or reset it (using VIC)... but never any of the BES servers (aside from the unused MDSS server that only has 512MB of RAM - JavaService.exe, or whatever its called, tends to spike in CPU after a while). does the CPU or RAM show excessively high in VIC when they're non-responsive to the remove sessions?

[BitNetMonkey]

Is this a fresh install or a P2V?

[TreeDude]

Can you RDP into the BES still? Our virtual console's CPU spikes sometimes and causes a similar issue (only it's all VMs, not just the BES). Usually it resolves itself after a minute and does not effect the VMs at all.

[yk101]

I have similar problems, but not with BES servers... Happens to the servers that experience heavy I/O load - Exchange, databases, etc.

This appears to be a VMware scheduler issue. I have a case open since 3.5 came out, but still no resolution.

If you move the VM to the local storage on ESX host it will run properly, but then you will not be able to use Vmotion and DRS.

[amadru]

I have pretty much the same setup (2 VM Bes on 3.5 with 500+ users each with a SQL 2005 backend cluster) and I have not seen any issues like what you are reporting. Is it just the BES servers that are crapping out? We have some phantom VM trolls that periodically make some of our 1200 VMs lock up, but no pattern.

Is your VC farm AMD or Intel? You using NFS or LUNs for your disk? Any VMotion going on?

[RadHaz75]

this happened to one of my vm bes after i first brought it online and was moving users over to it. i guess activating all those users and sending service books to them killed it. i found adding a second cpu fixed it, but brought it back down after all users were moved.