We are having an issue with repack of multple “nearing medium life” tapes from the same tape pool. I want to rule out CTA as the issue.
We are running repack on 3 different tapes from the same tape pool. The tape pool has plenty of supply tapes available and the repack vo is set to use 10 write drives. The read/retrieve has completed quickly. The archive/ write has been running on 1 destination tape for over a week. The CTA is only write between 1 to about 15 files per mount of this tape. It is very slow. There are over 30 k files that need to be repacked.
The libraries are not busy.
Why is the repack only using 1 drive at a time? Why is CTA only archiving to one tape a time? Why does it only write a few files for each mount?
Is something mis-configured on our end? Or is there some issue outside of CTA that is the root cause?
Thank you,
Jeffrey
Ok we think we found part of the issue. The tape pool said there was 233TB free space, but all the tapes except for 1 was set to full = true. After adding additional blanks to the pool, now those tapes are also being used for archiving.
I am still not understanding why only a few files are being written for each mount. We are using NFS as the buffer area, and someone on our team has suggested this could be slow and perhaps CTA is requesting the files, but the buffer is slow and timing out? We will investigate logs.h
Another update.
We had 3 tapes from the same tape pool added to repack. This may have caused our issues. The archiving was conflicting with files some how. We removed 2 of 3 repacks from the same tape pool, and now the archiving is proceeding normally with larger number of files per mount, and drive speeds near optimal.
Hi Jeff,
Thank you for the update, and I’m glad you managed to solve part of the issue.
Regarding the remaining problems, they could be related to NFS but with 30K files I don’t think that’s the problem… It’s a relativelly small amount.
However, without more information it’s difficult to tell.
Could you please share more information, such as any ERROR logs occurring around the same time?
The archiving was conflicting with files some how.
This could be some conflict in the CTA Catalogue. Can you share the error messages and/or any relevant logs?
Also, in which version of CTA are these errors being seen?
Cheers,
Joao
We feel this was NFS (buffer) related. One of the NFS nodes had issues. After those issues were resolved. Repacking resumed normal speeds, and multiple tapes from the same tape pool are not an issue.