cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Announcements
Field Notice 70545
2432
Views
0
Helpful
14
Replies
jim_nickel
Beginner

Cisco UCS Blades and vSphere DPM

I followed this guide:

https://supportforums.cisco.com/docs/DOC-8582

And it worked, but I have 2 problems.

1) The blade being put into Standby starts back up immediately - almost like a reboot -

and doesn't stay in Standby mode

2) The Blade being put into Standby mode has Faults on all vFabric's etc because it is

"off"

Any suggestions?

Jim

14 REPLIES 14
Robert Burns
Cisco Employee

Your issue is likely you have Power Management within the DRS  settings enabled.  If there's limited hosts, vCenter will immeidately  re-power up a host put into standby.  This is expected behavior for vCenter protecting the cluster.

Test - Disable DRS on your cluster, and then put a host into standby.  It should stay down indefinately.

Robert

Sorry...not the case.  We have 10 Blades and they are highly under utilized. Plus, DPM is in Manual mode right now.

I am testing and have not done a successful test...in other words, I have not been able to "Power On" because it won't stay down.

Jim

Shoot and a miss

I'd probably want to see the management logs from the host.  See if there's any IPMI calls coming from VC powering it back up.  Can you pull & attach a show tech for this blade (CIMC)?

Also need UCSM version, ESX build etc.

Thanks,

Robert

ESXi 5.0.0, 623860  - Nothing in the logs that shows VMware is sending another signal - only the Standby signal. In fact, it thinks it succesfully put it into Standby mode as the blade goes unresponsive for long enough.

BIOS: B230.2.0.1d.0.111620111513

Board Controller: B230100A

CIMC Controller: 2.0(1w)

One suspicion is that maybe UCS is interpreting the signal from ESXi as a reboot instead of a Standby.

If this is the case, then a firmware update is likely required.

Jim

I just tested this on my B200 M3 blades, works fine.

Let me test this on a B230 with that BIOS.

Provide the UCSM CIMC tech bundle, it will tell us if your suspicion is right or not.

Robert

Ok...here it is.....I think I picked the right CIMC ID - but I couldn't find the right spot to find the CIMC ID #, but it is chassis 1, blade 1, so I used CIMC ID 1.

Jim

Jim,

That's not the correct bundle.  (Sorry this thread got lost for a while).  That's a Mezz card show tech, I need the CIMC show tech.  You can grab this from UCSM.  The CIMC ID = the blade slot #.

Regards,

Robert

Hi Rob and Jim,

How did you guys progress with this one?  I am having the same issue and be interested to know the solution.

My environment is vSphere 5.0 with B230 M2 and DPM is not yet enabled.  UCS Manager and CIMC are running 2.0(1s).

On testing for Standby from the vSphere client, the blade reboots automatically (no power down).  A shutdown command gives the same result (which is a reboot) with a host connection failure alert.

Any help is appreciated.

Thanks,

Noli

Guys,

There's a known BIOS bug with EX platform hosts & ESX 5.  We're tracking it as Cscua63111.  It's tentatively slated for the 2.0(4c) patch coming out before the end of the year.

Regards,

Robert

Robert,

   Is there still a 2.04b coming out in November prior to 2.04c?  Or has it morphed into one release?

Kris

Yes, there will still be a 2.04b release coming out in the next month.

Robert

Should we wait for 2.04c or just go with 2.04b?   We need at least 2.04b for the SNMP fixes.

Is there a list of the fixes coming in 2.04c?

We don't use DPM, so that isn't any issue for us.

Kris

Kris,

I don't have a list of fixes going into 2.04c I can share.  Fixes are still being committed.

I would go to 2.04b if you need any of the fixes in it, otherwise if you can wait a couple months, wait for the next patch (2.04c).

Robert

Hi Robert,

I was wondering if this issue has been resolved i.e host going for reboot when tried to put it on standbymode.

 

Regards,

Shivaprasada.M.A

Content for Community-Ad