BladeSystem - General
cancel
Showing results for 
Search instead for 
Did you mean: 

Restart of OA blows up Virtual Connect

rbrewington
Occasional Visitor

Restart of OA blows up Virtual Connect


Whenever I reset my OA my VC ethernet gets kicked off the network for 30 seconds and sometimes my VC San gets disconnected. The only way to get the SAN back is to restart all the VC modules.

This is not cool.. anyone else ever have this happen? I have two OA's
11 REPLIES
Adrian Clint
Honored Contributor

Re: Restart of OA blows up Virtual Connect

Well your VC Console goes thru the OA NIC port...so if you reboot OA when you have only one OA it will kick it off.

Do you have compatible firmware levels for your OA, BIOS, ILO and VC Eth/SAN

http://h18004.www1.hp.com/products/blades/components/c-class.html#tab3_content
Check Compatability TAB!
rbrewington
Occasional Visitor

Re: Restart of OA blows up Virtual Connect

I have 2 OAs. All firmware is at newest versions. When I reset an OA my production Virtual Connect network goes down. And my VC SAN gets disconnected from the storage.
Batesydw
Occasional Visitor

Re: Restart of OA blows up Virtual Connect

I had the same issue on 3 blade enclosures I did firmware upgrades on.

I upgraded all blades to 11/08, then upgraded VC with auto update to 2.01. I didnt loose any servers as my TS session stayed up but my event log spammed with outages. I was testing a few things with teaming and updates.

I then upgraded the OA's to 2.41 and lost the VC manager totally and couldnt ping any of the VC's either. I had to reset everything.
WFHC-WI
Honored Contributor

Re: Restart of OA blows up Virtual Connect

I think what Adrian is referring to is connectivity to the Virtual Connect Management homepage. You would lose connectivity to this when rebooting a single OA (unless you have redundancy, which it sounds like you do.)

Out of curiosity, what method are you using to reset the OA?
rbrewington
Occasional Visitor

Re: Restart of OA blows up Virtual Connect

I realize that I will loose my management connection when I reset an OA. When I reset my OA it affects my production server network on virtual connect. It drops for about 30 seconds. And... one time it dropped all my SAN connections too.

When I reset the OA i just hit the virtual reset button
NJK-Work
Honored Contributor

Re: Restart of OA blows up Virtual Connect

Same problem here. In my case, I called HP tech support because two brand new BL495c G5 servers were not getting an iLO address from the OA. Tech support told me to pull out the OAs to reseat them. Against my better judgement I did this and ended up causing three of my blades to BSOD. All my servers boot from SAN. So I must have lost SAN connectivity for a time.

I have since set this registry setting to help alleviate brief SAN outages:

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Disk]
"TimeOutValue"=dword:00000078

Have not had the oppertunity to verify this helps yet.

Nelson
cjb_1
Trusted Contributor

Re: Restart of OA blows up Virtual Connect

Yes. We have this too. Same fw revisions.

We can pretty much repeat on demand by simply failing active to standby.
Only difference is the resulting problem.

So far we have seen -

New active OA fails to identify the first blade in the enclosure (comes up unknown and can only be brought back to life with a reset from the command line). Other blades fine.

Loss of san (as described above - reset all fc cards to fix)

Loss of OA - that's to say the active to standby fails to complete. Only way to resolve is to reseat original active OA.

We are looking into possible problems being network related as the later issue seems to flood the network.

Calls being raised with HP, will post results.
rbrewington
Occasional Visitor

Re: Restart of OA blows up Virtual Connect

I had to take my OA's back to 2.25 and then re-flash them to 2.41. Then take my VC modules to 2.02.

That fixed my issues. After I did this I was unable to reproduce the problems.

I also lost SAN and caused a network flood.

I had a case open with support but a post sales engineer actually suggested I do this. I told support what I had to do to fix the issue.

During the network flood my entire enclosure was powered down for a while. This could have also played a part in my problems being fixed.
cjb_1
Trusted Contributor

Re: Restart of OA blows up Virtual Connect

Thanks for the info.

Will give it a try. There's nowt in the 2.02 fw note to suggest a fix for this but there you go.

Been testing all morning and am not convinced the san is actually disappearing. Servers that are boot from san don't seem to be affected.

Connectivity through the OA and VC manager is obviously hampered by the network flooding across the management lan. In all but one of our enclosures this recovers after a while.

cjb_1
Trusted Contributor

Re: Restart of OA blows up Virtual Connect

We've nailed the loss of OA on connectivity (and network flooding)to settings on the cisco network switches(portfast).

The following was found in the vc 2.10 release notes:

Resolved an issue that the servers using FC might lose SAN connectivity for a brief period when the primary OA module was reset.

Be updating and testing this then...

rbrewington
Occasional Visitor

Re: Restart of OA blows up Virtual Connect

Did you disable portfast? Because thats what my network guys did to prevent the flooding from happening again.