Alletra 5010H Handover issue

[sangbae]

There are Main and DR storage. After handover in Main, rollback was performed. After rollback, the same symptom as the attached photo occurred. OS is Linux 8.10 version and multipath is set.

When viewed in multipath, only 1:0:0:0, 1:0:1:0 should be visible, but 1:0:1:100 1:0:0:100 volume is visible. Even if deleted in OS, the disk continues to be visible after reboot. Please analyze the cause.

※Before handover, the host volume was made inactive using umount and vgchange.

 

[buzzsubash]

Hello,

This needs some checks. Can you please tell me what is the version of NLT (linux tool kit) running on the OS?

[sangbae]

If you are referring to the HPE Nimble Storage Linux Toolkit (NLT) tool, I have not installed it. Is it required to be installed?

[buzzsubash]

Yes, that's the one — it can be downloaded from HPE InfoSight by clicking here . The NLT tool takes care of multipath settings, as well as iSCSI/FC adapter parameters like timeouts, queue depth, etc., based on Nimble best practices.

That said, handover/path issues like this typically require multiple log checks and environment-specific validation. To avoid back-and-forth and sharing sensitive data on a public forum, I'd recommend opening a support case.

Nimble/Alletra support is 24/7 — if you're able to call in, an engineer can set up a Zoom session and help resolve it faster.

Is the array under active support?

[NCG1]

Hi Sangbae,

Thank you for your email and attaching the screenshot. 

I wanted to know what type of replication is configured between Main and DR storage. 

Is it Synchronous (Live) or Asynchronous (Snapshot) replication?

How was the handover done on the volumes? Can you please let me know what documentation you followed?

Please let me know if you have any questions.

Regards,

Naveen Chandru G

[Hdcom]

Hello, this story uses the FC connection method, not the ISCSI method. Do I need to install NLT?

[buzzsubash]

Hello @Hdcom 
Yes, NLT is  recommended for both iscsi and FC setup. It does set recommended multipathing policies and lot of host tuning, which is otherwise to be done manually and is prone to miss out few.

We strongly recommend to install NLT on all the hosts where nimble volumes are connected.  Here is the link to our host integration guide  https://infosight.hpe.com/InfoSight/media/cms/active/public/pubs_Linux_Integration_Guide_4_0_2.whz

[NCG1]

Hi Sangbae,

Hope you are doing good.

Please contact our support team via phone call from the numbers in the attached picture or email to them using support@nimblestorage.com. 

This needs troubleshooting and analysis from the support team.

I will go ahead and close this ticket.

Please reply back if you need any further assistance.

Regards,

Naveen Chandru G

[sangbae]

I have an additional question. I found a volume corresponding to that LUN. That volume is a snapshot volume. How can I delete it and why the LUN was assigned to the snapshot volume? And to erase the origianl snap, can I just delete it from the snap collection?

[buzzsubash]

LUNs does not get assigned to snapshot volumes.
Are you using Veam for backups or VMWareSRM for DR ? Veam or third party backup tools make the snapshot onine inorder to take backup and make them offline once itis done. We have seen cases where they are left online for many reasons.

Also, when setting ACL, only the volume should be accessible to the host, in this case I could see ACL is set to Volume & snapshots. if you are using Veeam in the environment, recommended setting is it's intiator ACL should be enable to access only the snapshots and for hosts just the volumes.

Please do double check from the host before deleting/offlining the snap volume that it is not being used for any IO. Once all checks are done, you should be okay to offline the snap from array UI, followed by an HBA/filesystem rescan on the host.

[sangbae]

I am not using Veem or any other backup solution. The host is SAP. I am also wondering if deleting that snap will recreate the full snap.

[buzzsubash]

Thanks for letting me know.

Looking at the snapshots list, Main-Dr-15min-Main-Dr-2025-06-09:02:15:00.000 of size 104013631488 (13 TiB) looks to be the seed snapshot and the time schedule of 15 min follows for the preceding ones. And if you look at the date of immediate previous one, it says 22-May-2025, whereas the above snapshot is from 09-June.

Deletion of the snapshot with name 'original' should not create a new one in this case. But, we need to perform few checks to make sure it won’t impact.

BTW, I can see you have already created a support case and my colleague is assisting you. You should consider calling the support line and we can do Microsoft Teams meeting if Zoom does not work. If you are looking for immediate resolution, a remote sharing is a must! Resolution over email will not work for such complex issues where data integrity/loss is at stake.

With the array in dark site and no access to logs/array, it is expected to be delayed. There is nothing much we could do if customer can't co-operate with support.

I hope you understand, please explain to the customer and arrange a remote sharing session.

[sangbae]

Thanks for the quick reply. I'll discuss the remote session with you.