S3 growth after svMotion to new production storage

[EWMarco]

Hi everybody

We're a bit confused with what we're currently seeing on our StorageGRID S3 installation.

We are moving production storage from our old A700 NetApp storage to a new A800 system using storage vMotion. We've seen jumps in our S3 fill state after migration days.

We are using a scale-out repository containing S3 buckets. The system is set to copy to capacity tier immediately and move after a certain amount of time.

We do not see a growth in space usage on our performance extents, only on the capacity S3.

To be honest, we have no idea why S3 is getting fuller at the moment. It is possible that there is no causality between the migration and the fill state of the S3 of course, but we're unsure.

The performance extents are ReFS or XFS. The backup is leveraging storage snapshots.

Can anyone think of a set of circumstances given our environment that would make it plausible that the migration is causing additional blocks to be written to S3?

Thanks

[david.domask]

Hi Marco,

By chance did you notice if new morefIDs were perhaps assigned to the VMs after the migration? To my knowledge storage vMotion should not do this, but easy way to check is to look at the backups under Backups > Disk, and see if the VMs that were migrated now have duplicate entries in the backup -- this likely means that the morefID changed, which would result in a new backup chain and explain the higher S3 usage.

If that does not check out, please open a Support Case and allow Veeam Support a chance to review the debug logs and understand the situation better. Be sure to include logs for Support to review. (Use the 3rd radio option to export from a given host, and select the Veeam server itself -- please note the name of one of the jobs backing up the migrated VMs in your case to help Support identify which job to focus on)

Please share your case number once created. Thanks!

[EWMarco]

There are no new morefIDs.

And I was so very afraid you would propose opening a case.

[david.domask]

Thanks for the confirmation Marco.

And regrettably, we won't be able to effectively troubleshoot over the forums, so a Support Case is the way to go. New Morefs were one "low-hanging fruit" so it was worth a shot, but a case will be the next best step.

[EWMarco]

Yes but the work involved in getting tens of gigabytes out of our environment to Veeam and then conveying the issue to support are quite staggering psychologically speaking.