Feature Request: SOBR rescan options

[markhensler]

I would like to be able to rescan individual SOBR extents rather than being forced to rescan the entire SOBR each time.

We have a very large COS bucket, this takes forever to rescan. If I have reason to rescan my local extents, I would like to be able to omit our COS storage in that rescan task.

Even more narrowly, if I have deposited some backup files on an individual extent, I would like to be able to rescan that one extent rather than all extents in the SOBR.

If this functionality exists, I've missed it. I have tried to right-click on extents under "Backup Repositories" as well as when viewing the SOBR under "Scale-Out Repositories". In neither location does the "Rescan" option appear.

[PetrM]

Hi Mark,

Thanks for the idea! However, there is one detail which might complicate implementation: the rescan gathers information about backups that are currently available in the repository and updates the list of backups in the configuration database, for example it can import backups. In case of SOBR, different fragments of the same backup can reside on different extents. Therefore, it's not clear how to handle import or update of information about backup in database when just one extent is being rescanned: a significant part of data can simply become unavailable.

On the other hand, I will appreciate it if you share more details regarding your problem related to slow rescan with us: how many extents do you have, what is the total size of used space and how long does it take to rescan the repository?

Thanks!

[markhensler]

Hi Petr,

In one SOBR, we have 7 local extents and 1 COS. The COS currently has about 145 TB of data. The last time this SOBR was rescanned, the COS took about 2.5 hours.

[PetrM]

Thanks for reply! May I ask you to raise a support request and to share case ID with us? I believe debug logs analysis might be helpful to look for a potential bottleneck in rescan process.

I forgot to clarify an important detail: how many backups do you have in total on SOBR and how many backups were updated or imported during the last rescan which took 2,5 hours? For example, according to our internal tests an import of single encrypted backup having 80 restore points and 400 VMs can take up to 20-30 minutes.

Thanks!

[markhensler]

Is this what you are asking for?

Code: Select all

9/16/2020 11:57:55 PM          Starting backup repositories synchronization
9/16/2020 11:57:55 PM          Enumerating repositories
9/16/2020 11:57:55 PM          Found 1 repository
9/17/2020 12:00:16 AM          Processing performance tier extents of DAL10 Storage
9/16/2020 11:59:27 PM          JZVeeam-BU6 Local-6a: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 28, failed 0
9/16/2020 11:59:14 PM          JZVeeam-BU3 Local-3a: added 1 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 0, failed 0
9/16/2020 11:58:54 PM          JZVeeam-BU3 NFS-35: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 0, failed 0
9/16/2020 11:58:57 PM          JZVeeam-BU6 NFS-33: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 1, failed 0
9/16/2020 11:58:54 PM          JZVeeam-BU6 NFS-32: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 0, failed 0
9/16/2020 11:58:57 PM          JZVeeam-BU6 NFS-30: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 1, failed 0
9/17/2020 12:00:13 AM          JZVeeam-BU6 NFS-31: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 49, failed 0
9/17/2020 2:28:27 AM          Processing capacity tier extents of DAL10 Storage
9/17/2020 2:28:23 AM          DAL10 Archive to US East: added 0 unencrypted, added 0 encrypted, updated 0, removed 0, skipped 78, failed 0
9/17/2020 2:28:28 AM          Backup repositories synchronization completed successfully

[PetrM]

Yes, it's exactly what I've asked to provide. Please don't forget to share the support case ID as well.

Thanks!

[veremin]

We're constantly working on improving performance of object storage rescan. So, QA team is interested to see your logs in case they've some issues or bugs. Thanks!