BladeSystem - General
1748185 Members
4583 Online
108759 Solutions
New Discussion

Can the Onboard Administrator tray be unseated without interrupting Virtual Connect operation?

 
chuckk281
Trusted Contributor

Can the Onboard Administrator tray be unseated without interrupting Virtual Connect operation?

Lawrence had an interesting customer question:

 

********************************************************************************************************************

Hello Team,

 

My email title is actually a loaded question.

 

I have read quite a number of posts that suggest that the OA tray can be reseated, and the only thing that happens is that the fans go to 100% speed.

 

I had a situation where both the primary and secondary OA modules in a customer’s C7000 enclosure were no longer responsive and so the 16 blades in the enclosure could not be managed via the OA even though they were running.  So whilst the customer could get a remote console through Microsoft Terminal Services, they could not use iLO on their BL460c blades because of the hung state on the OAs.

 

My customer asked me what they should do.  I recommended removing the OA1 then OA2 then reseating the OA tray and then reinserting OA1, wait until it was working and they can login,  and finally reinserting OA2.

 

They asked me if this was safe to do, and I’ve seen lots of posts on this forum where the recommendation was to reseat OA.

 

Unfortunately for this instance, when the OA1 module was reinserted, this caused the VC communications to drop, as if the VC modules (both Ethernet and FC) went through a process of resetting themselves.

 

From the VCM login screen, the checkpoint symbol was visible.

 

This customer is running original VC Ethernet and FC modules, on VC firmware 1.32 and the OA is on 2.51, but they are about to embark on upgrading their fleet of enclosures to the latest Firmware Set – they do their upgrades every 12 months.

 

So the question remains – is it considered “safe” to pull the OA tray from a production c7000 and expect the VC modules to keep running, under normal circumstances?

************************************************************************************************************

 

Troy had an answer for Lawrence:

*************************************************************************************************************

There is a fix in VC firmware 2.10 for interruption to VC FC operation that can be observed when primary OA is reset.  I’m not aware of any issues that would cause interruption to VC Enet operation though.

******************************************************************************************************************

 

Have you tried this? Let us know your findings.

2 REPLIES 2
jlackman
Occasional Advisor

Re: Can the Onboard Administrator tray be unseated without interrupting Virtual Connect operation?

Under normal circumstances, yes the OA TRAY should be able to be reseated with zero impact. Of course, all of the customary disclaimers should go here now (disclaimers etc.). I have had HP support tell me this straight from their guys, and we've done it in the past. But in a situation like you're descibing, who knows since there seems to be something else going on as well to create an abnormal behaviour.

 

Jonathan


@chuckk281 wrote:

Lawrence had an interesting customer question:

 

********************************************************************************************************************

 

Hello Team,

 

My email title is actually a loaded question.

 

I have read quite a number of posts that suggest that the OA tray can be reseated, and the only thing that happens is that the fans go to 100% speed.

 

I had a situation where both the primary and secondary OA modules in a customer’s C7000 enclosure were no longer responsive and so the 16 blades in the enclosure could not be managed via the OA even though they were running.  So whilst the customer could get a remote console through Microsoft Terminal Services, they could not use iLO on their BL460c blades because of the hung state on the OAs.

 

My customer asked me what they should do.  I recommended removing the OA1 then OA2 then reseating the OA tray and then reinserting OA1, wait until it was working and they can login,  and finally reinserting OA2.

 

They asked me if this was safe to do, and I’ve seen lots of posts on this forum where the recommendation was to reseat OA.

 

Unfortunately for this instance, when the OA1 module was reinserted, this caused the VC communications to drop, as if the VC modules (both Ethernet and FC) went through a process of resetting themselves.

 

From the VCM login screen, the checkpoint symbol was visible.

 

This customer is running original VC Ethernet and FC modules, on VC firmware 1.32 and the OA is on 2.51, but they are about to embark on upgrading their fleet of enclosures to the latest Firmware Set – they do their upgrades every 12 months.

 

So the question remains – is it considered “safe” to pull the OA tray from a production c7000 and expect the VC modules to keep running, under normal circumstances?

 

************************************************************************************************************

 

Troy had an answer for Lawrence:

 

*************************************************************************************************************

 

There is a fix in VC firmware 2.10 for interruption to VC FC operation that can be observed when primary OA is reset.  I’m not aware of any issues that would cause interruption to VC Enet operation though.

 

******************************************************************************************************************

 

Have you tried this? Let us know your findings.


 

Server-Support
Super Advisor

Re: Can the Onboard Administrator tray be unseated without interrupting Virtual Connect operation?

What about resetting the OA module ?

I assume that when updating the OA firmware, there will not be any outage as well for the production BLade servers.

Best regards,