This post is more than 5 years old
2 Intern
•
326 Posts
0
1652
August 25th, 2013 18:00
Reclaim space from Tape ?
7.6.3.7
Hi all
The tape marked full with around 2.7TB on LTO5.
There were ss on tape with aborted status. Manually expired them expecting to regain some space back but didnt.
Mount and unmounted in hope it "refreshes" but same. Content just 500GB but showing full.
Very rarely this happens and do the expiry manually so I may have forgotten some pointers...:-)
Any idea how to reclaim the space back?
Thank you
No Events found!
masonb
445 Posts
0
September 3rd, 2013 01:00
Thierry101,
Space cannot be reclaimed from tape for aborted, expired savesets as we cannot fill in the gaps which would be left. Tape technology is sequential – the only way to reclaim space once a tape is genuinely full is to re-label. If a tape is marked as full prematurely you can reset the status to appendable, however in your case as it has reached the end of tape (EOT) marker at the end of the physical tape this is not possible.
You will have to wait now until all savesets are expired or you could stage the active savesets to another volume and then re-label/recycle this one.
Regards,
Bill Mason
TimQuan
6 Operator
•
1.2K Posts
0
August 26th, 2013 22:00
“Tape marked full” issues can be caused by many factors. Usually we can:
1. Verify that the tape drive(s) are not dirty. Perform drive cleaning to ensure drive head performance.
2. Verify that there are no I/O errors in the daemon.raw file, and operating system logs (Event Viewer, /var/adm/messages, etc.)
3. Verify that the media database on the NetWorker server is not corrupted.
4. If the drive is accessed via a SAN HBA, ensured HBA firmware and drivers are up-to-date.
5. Ensure that the OS accesses the tape drives using variable block size. This is done via the st.conf file on Solaris, or equivalent method on other OS's (AIX-SMIT, HP-SAM, Windows-registry).
6. Ensure that the OS is up-to-date on tape related patches, along with SCSI/SAN patches (pending on how the drives are accessed).
You'd better open daemon log to check the detailed error message. That will give us clearer information what's the root cause.
o17Uu33DCF12520
6 Operator
•
1.1K Posts
0
August 26th, 2013 23:00
how many tapes have this problem? and how many drives are involved?
try another drive and see what's gonna happen. if it only happens to a specific drive, you better check the drive status and any hardware error.
please mark my answer as "correct/helpful answer" if it helps.
ble1
6 Operator
•
14.4K Posts
•
56.2K Points
1
August 29th, 2013 02:00
Unlike disk, you can't reclaim space on tape by removing backup sets (at least with current generation of tapes). However, if your tape is marked full after 500G, you can try to change its status to appendable and see if you can write more data onto it. If not, that tape is most likely bad (you can verify it by relabeling it once all backup sets are expired and checking when it gets full next time). What you can do is nsrstage ssids from one ("bad") tape to another ("good") and that way get "bad" tape in state where you can relabel it afterwards. Of course, you may wish to check previous entries in logs to check when tapes are marked full and see if there is any pattern (like same barcode or always the same drive or same time of the day when it happens).
Thierry101
2 Intern
•
326 Posts
0
September 2nd, 2013 13:00
Hi All
Let me rephrase...the drives are not bad....the LTO5 tapes gets full between 2.7TB to 4TB depending on data type...
IF some save sets are to be removed from the tape, would it free up space on tape? in another word, would the status change from full to say 30% used and can be used by next backup/clone ...?
Thanks folks