ViPR with replicated Isilons and Superna Eyeglass

Hi Terence,

There is no known issue with vipr controller and Eyeglass.  Can you clarify what does not work.  Eyeglass can sync VIPR created shares or exports without any issue.  The location in the filesystem is /ifs/vipr......   so you need a syncIQ policy created to protect the VIPR data.

Can you explain how this is setup with SyncIQ.

We have not tested VIPR controller explicitly but I don't see anything that would not work.  The shares and exports created by VIPR are not special in anyway.

Regards

Andrew

Hi Andrew...

Well that is strange. The last statement we got from an EMC SE was, that it´s an actual show stopper. We did not yet set up the Isilons in any way. The Superna Eyeglass Installation is scheduled for monday and from there, we will continue to deploy the rest of the config and the file migrations. I will double check the statement we got from EMC and who actually said that it wouldn´t work in the first place. Will get back to you as soon as possible.

Greetings and thank you for your reply

Terence

if you can get the technical answer from the SE.   I see no technical issues at all, and would like to understand them if there are any.   If its just a "not listed on the matrix" response that is different.

Regards

Andrew

shares won´t "show up"...  They will but only requirement is a SyncIQ policy is created to protect the VIPR shares and exports (a single policy could be used to protect all VIPR shared data on Isilon.

Other than the above  all standard failover features are supported DFS mode, Access zone failover for VIPR created data.

Regards

Andrew

We are waiting for the statement from EMC support right now. It might be related to the "not tested" statement in case of support calls.
One statement I got already was, that ViPR created shares won´t "show up" in Superna and Vice Versa. But anway. Let´s see what the EMC ViPR Support says.

Here is the results of quick test that shows Eyeglass protects VIPR configuration without any issues.

Setup was the following:

1. source cluster onefs 7 with vipr 3.5 installed and it created shares and exports
2. target cluster onefs 8 (to show Eyeglass can protect vipr even when source and target clusters are different versions)

DR cluster before Eyeglass protects

Eyeglass job showing vipr configuration data to be protected

Eyeglass job running and creates shares and export on DR cluster (onefs 8)

DR Cluster after eyeglass job runs showing newly created vipr configuration

At this point failover is normal eyeglass failover access zone or DFS mode would work fine to failover the vipr configuration (SPN's, dual delegation zero DNS, quota failover, etc....).  So from a DR view.  No issues to protect Vipr configuration data and failover

The only thing I can not answer is if the source site and VIPR install are destroyed, how does new vipr installation at the DR site needs to inventory , and ingest the existing configuration data that is now writeable on DR cluster.

This set of procedures should already exist for VIPR customers, I have not looked for it.  VIPR can ingest inventory existing clusters.  This would be normal VIPR DR recovery procedure at this point  since all data and configuration is failed over by Eyeglass.

I suggest sending this thread to VIPR team to review.  They can contact me as well if needed.

Andrew

Thank you very much Andrew for all the work you did to write this. I forwarded this article to the EMC team, let´s see what they say.

Regards

Terence

I posted the solution description on our website.

http://documentation.superna.net/eyeglass-isilon-edition/eyeglass-solutions/vipr-data-protection-with-eyeglass

Did not yet get an answer but I´m still tracking the topic. I´ll keep you informed about it ASAP.

The checks will take approximately six weeks for the EMC team to finalize every test. If it´s finished earlier, I´ll get back to here immediately.

Thanks for your support Andrew.

we are in contact with VIPR engineering now on this testing.

Regards

Andrew