Enhancement Request: Parallel Tape Cataloging & Multi-Server Processing

[olafurh]

Hello Veeam Team,

When I initiate a tape catalog task in Veeam Backup & Replication, it processes tape one at a time - using only a single tape drive - even if the library robot has multiple drives available. On libraries with hundreds or thousands of tapes, this serial approach creates a significant bottleneck and can take days to complete.

Proposed enhancements:
Allow the catalog task to distribute tape catalog process across all available drives in the robot, so that multiple tapes can be scanned and cataloged concurrently.

Distributed cataloging across tape servers? :) And event tape jobs.....
Enable a single partition to leverage more than one tape server please, balancing tasks across servers to further shrink processing windows.

I believe these enhancements would make Veeam’s tape cataloging far more efficient for large environments.

I’d welcome any interim workarounds!

Oli

[david.domask]

Hi olafurh,

Thank you for the request -- the cataloging process indeed needs some focus on this element as the situation is understood, it can take a long time on current generation tapes, so parallelization would indeed be useful.

Can I ask, how often are you doing bulk cataloging? Approximately how many tapes? Cataloging in my mind should be fairly normal but uncommon event, but would be curious on what the workflow is that results in frequent cataloging. The request makes sense, just trying to understand the use case.

>Distributed cataloging across tape servers? And event tape jobs.....
Enable a single partition to leverage more than one tape server please, balancing tasks across servers to further shrink processing windows.

For cataloging across tape servers, I guess we need parallel cataloging first, but makes sense. As for tape jobs however, you can already configure multiple libraries for a media pool; depending on your configuration (active:passive or active:active), you can have a job be processed by multiple tape libraries attached to different tape servers. I advise review the User Guide page on Parallel Processing to understand how the media sets and tape selection will be handled.

[olafurh]

We don't often need to catalog tapes - in fact, we've never had to before, and hopefully never - but if we lose our entire database or environment due to a major incident, or need to ingest tapes from another setup, cataloging becomes critical. We tested this scenario and discovered the process was surprisingly slow, even though we had plenty of tape drives available. In a disaster recovery situation, quickly importing and restoring petabytes of data would be essential, and cataloging speed would become a serious bottleneck.

The current limitation of having only one tape server per partition restricts throughput, as tape drives must be divided across multiple partitions. If Veeam supported running multiple (2–10) tape servers per partition, we could leverage 20–40 drives simultaneously, dramatically boosting performance. Right now, we limit our setup to 8 drives per partition, handling about 800 tapes per partition, and we're maxing out around ~25 Gbit/s throughput. Increasing parallelism would significantly accelerate this critical recovery process.

Oli