[Tom Goldberg]

Kyle,

On Cache-A units, the catalog is not used when reading tapes, it is only there so humans can know what is on each tape. So importing your backup catalog wouldn’t help (even if it were possible, which it no longer is).

Any reading of a tape uses the table of contents on that tape to access files. If the loaner Prime-Cache can’t read it, either there is a problem with the unit or with the tape. Following are some things to try.

If the tapes are written with LTFS, you might want to try a more up-to-date LTFS system to recover them. If they are written as tar tapes, there is a procedure available to recover all files from any Cache-A tar tape – you can find that here:
https://www.tomgoldberg.net/Cache-A/CATechBrief-ReadCAtar.pdf
also a set of instructions for using terminal commands on a Cache-A is here:
https://www.tomgoldberg.net/Cache-A/CATechBrief-CommandLine.pdf
and if you did want a copy of the catalog, these instructions will work on your old Pro-Cache assuming it will boot:
https://www.tomgoldberg.net/Cache-A/CATechBrief-catalog-export.pdf

I hope this helps,
Tom

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Ed, that is strange behavior indeed.

First, I suggest you turn off tape spanning, at least until the issue is resolved. To look into this, I’d need to know if you are having the issue with both LTFS and tar, and what the catalog and tape info say about these tapes that are ejected.

I suggest you generate a log file (Diagnostic Summary tab) and send this info to me directly at the address below and I’ll do what I can to help you.

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Hi Lee

I’ve never seen that warning, perhaps you tried this with your server mounted? Disconnect any source volumes.

The command you should run is fsck -C /dev/md0

If that still gives you such a warning, or if you need further assistance, please contact me by email off-forum and I’ll try to help you out.

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

In fact any linux or unix machine with an appropriate mounted LTO drive can read Cache-A tar tapes.

The general form is to issue the command Jack wrote in his previous post, although this can be automated with the command:
while [ 1 ]; do tar -xvf /dev/nst0 -b 300 -C /media/vtape/restore_directory/; done
Full instructions for doing so may be found at Reading Cache-A Tar Tech Brief.

I would note that if a Cache-A unit serves your environment (as it sounds like it does for Lee) there is no reason to stop using it just because ProMax is dropping it. Errors occurring during backups are due to correctable causes:

Make sure your source drive(s) have a reliable route to the Cache-A – sometimes errors like these are just a bad ethernet cable or switch

Run a cleaning tape to eliminate the possibility that this is just due to dirty heads

Run fsck on the Cache-A file system to detect and correct file system errors. You can also run the smartctl commands to check on the health of your disk drives. (any linux guy can help with these or contact me for instructions)

Run HP tape tools analysis to check on the health of the tape drive (installed on every Cache-A – run by issuing the command /opt/ltt/hp_ltt) – replacing the tape drive is much less costly than buying a whole new system

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Mike,

Assuming your Cache-A works with other LTFS volumes, this tape is bad. If all your LTFS tapes do this then it’s the system but I expect this is the only one. From your terminal output, it looks like the tape is empty in addition to not mounting. Something damaged it beyond the ability of the existing LTFS tape tools at the time to fix it. My only thought that might help would be to try to repair it with newer LTFS software (i.e. a current copy of PreRoll Post or Yo Yadda).

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

You undoubtedly have “type C” video tape there. It will play back on a variety of Ampex and Sony decks listed on Wikipedia under that topic: https://en.wikipedia.org/wiki/Type_C_videotape

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Hi Randy,

Unfortunately the old Cache-A authorization servers are long gone. ProMax is the only place to go for this problem.

Sorry about that.

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Unfortunately there is no way for a user to restore the database from backup (we never did finish developing that capability) – the only folks who can do that at this point in time is ProMax support.

It still sounds like it might just be too much data in the library, but alternately you may possibly be having HDD issues.

You may contact me at my email address below for some more detailed ideas about how you might best move forward with the Cache-A. Let me know what model and version of software you are running.

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Hi Jack,

Your slow performance can be due to a few things – possibly tape drive issues but more likely an overly large number of files on this tape or possibly your tape catalog has grown so large it it taking too much time.

LTFS itself gets bogged down in any archive where content is stored as individual frames. If you think this might be the problem, try doing a tar archive on the content instead and see if the performance is better.

If you think it is might in fact be your tape catalog, you can try deleting some older tapes from the catalog if you don’t need to track them anymore or can afford to track them externally (how to export your catalog is in a document linked below). You can also try a database repair by logging in to the unit for a terminal session as described in the other document linked below and execute the following commands:
(Commands must be in order – its a good idea to backup your catalog first)
/etc/init.d/httpd stop
/etc/init.d/tapemgr stop
/etc/init.d/mysqld stop
cd /media/catalog/database/mysql/catalog
myisamchk *.MYI
myisamchk --recover ‘toc’
myisamchk --recover ‘inode’
myisamchk --recover ‘file’
myisamchk *.MYI
/etc/init.d/mysqld start
/etc/init.d/tapemgr start
/etc/init.d/httpd start


If you think it might be the tape drive itself, you can always run a cleaning tape pass which might help. Or if you want to actually check the health of the tape drive, the full suite of HP tape tools are on your system – you can run them from the command line by typing /opt/ltt/hp_ltt Instructions for using these tools can be found in HP’s online documentation but it is menu driven and you may be able to figure it out by poking around.

Instructions for logging into the Cache-A’s command line can be found at
https://tomgoldberg.net/Cache-A/CATechBrief-CommandLine.pdf

The Cache-A tech brief describing how to export your catalog is at:
https://tomgoldberg.net/Cache-A/CATechBrief-catalog-export.pdf

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net

[Tom Goldberg]

Darren,

I’m sorry to advise that the Cache-A won’t automatically rebuild a failed RAID onto a new unformatted drive, it needs to be specifically partitioned for the job.

Please contact me off forum at the address below for further help.

Regards,
Tom

Tom Goldberg
TGCS
30201 Rainbow Hill Rd.
Evergreen, CO 80439
mailto:tomgoldberg@gmail.com
https://tomgoldberg.net