Check status of the svt-zone-realize command

[J-Philippe]

Hello, is there a way to control how is going the "commit" of the availabilty zone when executing the svt-zone-realize command ? For instance when adding new nodes in a federation after assigned the nodes to the zones. In my case the new nodes are still displayed this way when running svt-federation-show (none)=> Zone name connected against the "legacy" node that are "Zone name" connected Like if the realize command was not successfull or still working (but not so many data to move to make the few VM compliant with the zone…) Regards, Jean-Philippe

[DeclanOR]

Hi @J-Philippe 

Thanks for your query.

Question: What version of OVC software are you currently running?

Am I correct in saying that a stretch cluster already existed, and that you have deployed new nodes to the fed and are trying to add them to the existing stretch?

Thanks,

DeclanOR #I am a HPE Employee

[J-Philippe]

Hi Declan, running the 3.7.7.178 version. You 're correct I've added 2 nodes to an existing stretched cluster. Deployment works fine (all green) for both nodes. To give you more info : One node seems to be unstable (case in progress with the HPE support) sometime loosing datastore access or from is own ovc considering some other nodes faulty that are not from all the others ovc. (all ping works fine mgt, storage, federation…) Other exemple is the support-capture that take too long to complete and asking if you want to wait or not… Regards, Jean-Philippe

[DeclanOR · Accepted solution]

Hi @J-Philippe 

Thanks for the info.

I would recomend resolving any issues before adding them to the stretch. I suspect a possible MTU issue by sounds of your explanation...but don't quote me I'm sure the engineer will resolve.

Once the environment is healthy, re-add the new hosts to the zones, and re-commit/realize.

If it's still the same at that stage, update the thread....or possibly create a seaparate case with the engineer to review.

Thanks,

DeclanOR

[J-Philippe]

I will share the resolution on this post. I think you are right regarding the MTU. "Legacy" nodes mgt ports have been moved from the 1Gb NIC to the 10Gb NIC a few month ago. So the MTU is still at 1500 on 3/4 except one at 9000 (don't know why) When I've deployed the 2 new nodes, I've set the MTU to 1500 for the management. However, the management in the vmware is using automatically the 9000 MTU (mgt ports are on the 10Gb NIC). All switchs ports are "Jumbo Frames ready". Maybe it can be something to go on… Regards.

[DeclanOR]

Sounds like it is very likely MTU yes.

Let us know how it goes once resolved, and re-added to zones/committed. It's possible that MTU issue is preventing the successful zone realization. 

Thanks,

DeclanOR  #I am a HPE Employee

[J-Philippe]

Sir Declan, you're right ! I've check the switch and the port regarding the node was not activate for the Jumbo frames. I've change it and the datastore are now reachable. Waiting for HPE support returns for the svt-realize but I will probably have to play it again. Keep you update. Thanks for the hint ! Regards, Jean-Philippe

[DeclanOR]

Hi @J-Philippe 

Glad to hear the underlying MTU issue was resolved. Were you now able to assign the hosts to the stretch?

Thanks,

DeclanOR

[J-Philippe]

Hi Declan, not yet, I'm waiting for HPE support return to confirm me all is fine. Yesterday the 2 new nodes were still like joigning the strech via the availablity zone. (None => Zone Name)

Regards

[DeclanOR]

Hi @J-Philippe 

Hope all's well.

Just checking in to see if you got this resolved afterwqards?

Thanks,

DeclanOR