Start a Conversation

Unsolved

This post is more than 5 years old

1785

May 17th, 2016 03:00

VPLEX metro upgrade 5.1 to 5.5

Hi Vplex community,

I have Vplex Metro  5.1 cluster (2 DCs, storage provisioned from VNX and VMAX ) with os ver. 5.1.0.02.00.01, I need to upgrade it to 5.5.

I need to assess that all hosts attached to vplex will be fine with the new version of Vplex.

What exactly schould be chcecked before that upgrade from hosts perspective ?

Are there any requirements in terms of hosts and soft running on hosts (OS, agents..etc etc.) for 5.5 which differ from 5.1 ?

I expect that that kind of upgrade should be completely transparent from hosts perspective but need to make sure if there is any risk for hosts.

What is the emc recommended upgrade scenario for vplex 5.1 to 5.5 ?

Kind Regards

Exp

15 Posts

May 25th, 2016 13:00

Hello,

We recently upgraded our VPLEX to 5.5. When scheduling the upgrade, EMC will ask what switches you use, and what code levels they are at, to the check the compatibility matrix.

It's my understanding that as long as your VPLEX can communicate with the switch, the servers will be fine (there connections to the switch are not changed).

However, make sure that all of your hosts have redundant paths to the VPLEX, as each director is rebooted during the upgrade.

I recommend opening an upgrade SR and verifying your environment with your local EMC resources.

Thanks,

Tanner

June 2nd, 2016 05:00

Hi Tanner,

Thank you for your reply.

All issues on the level of switches are already addressed. SR was raised earlier and CCA process was initiated (it's approved on emc side for 30 days when you raise upgrade SR), however later on upgrade was canceled by client but issues found during CCA process from grabs or on switches level were fixed.

Currently client wants to ascertain beforehand whether all initiators on the VPLEX will be compatible with version 5.5, I'm trying to archive that by SR raised for host team to analyse esxi, linux, wintel grabs and check it through Elab Navigator tool.

Have you had any issues in terms of initiators compatibility ?

Kind Regards

explorer

15 Posts

June 3rd, 2016 14:00

Hi Explorer,

We did not have any issues with compatibility in any of our environments. It may depend on which exact version of 5.5 you are going to, but one thing that did catch us is the following new feature (from release notes):

  • Local storage volume latency monitoring

VPLEX monitors the latency for reads and writes that complete to a storage volume locally on an array. VPLEX samples the latency every five minutes by default, and they are measured against a set of performance policy for the cluster. Storage volumes are marked degraded if the latency values exceed the set threshold. Else, the volumes are marked healthy. This feature is enabled in VPLEX by default.

As soon as we upgraded the management console, LUNs were being reported as 'degraded'. We thought something had gone wrong. It turned out that the high latency on the LUNs was pre-existing, and that the VPLEX could just report it back to us now. We proceeded with the upgrade, and completed it without any issues.

I hope that helps,

Tanner

No Events found!

Top