Optical disks in 3995 library 2

[J1gh2]

Hi Folks

Would anyone know how to checkin optical disks into a 3995 library without having to check them in through TSM. We have to recover old data from some disks that need to be checked into a spare library that is not attached to TSM.

Thanks

 

[Yrrk]

It is not possible to read TSM created media without TSM.

 

[Yrrk]

... You will either have to attach the spare library to your TSM server and check them in, or install TSM on the server that has the library attached to it and recover your TSM database backup on that server, check in the media, and start reading the discs.

 

[J1gh2]

ok. Thanks mate

 

[J1gh2]

Yrrk

I found that there is a way to checkin these optical disks from the AIX command line (and without going through TSM). One of our UNIX guys used a little command line based programe (called lbtest) that interfaces with the 3995 library and we were able to check in the disks. This actually worked much faster than the normal checking process of reply with process numbers etc.

Just thought you may want to know.

Cheers

 

[Yrrk]

Are you able to actually read the disks without TSM? Or what exactly do you mean by checking in the disks?

 

[J1gh2]

Yes, we are reading data from the disks from a script outside TSM. We checked in the optical disks into the 3995 library (without using checkin libv command from TSM). However, this is an extremely rare recovery scenario so you may not have come across it before.

Essentially what happened was that we decommissioned some nodes and then found that we needed to recover data from those nodes but couldn’t just register the nodes again and then recover since the server doesn`t know anything about them anymore. So we contacted TSM development team and they have developed a script that will checkin the disks and read the contents of the disks and then rebuild the database so that we can then recover the data. So the script is working but it’s extremely slow (takes about two weeks to read and reconstruct ten 5GB disks). We are not reading the disks from TSM but from a script.

Hope that’s clear.

Regards

 

[Yrrk]

This is really interesting.. So you will ultimately recover the data using TSM after reconstructing the database with said script? i.e. the script itself isn't allowing you to restore the data or even "read" it per say.. TSM is still involved?

Any thoughts as to why it takes so long to run? That seems far slowerly then drive speed.

Know if it works with other media as well like tapes? i'm guessing so?

 

[J1gh2]

You are correct. The script is reading the optical disks and copying the data to ESS. After this is all done, we will reconnect the ESS back to TSM, re-register all the nodes that were deleted, and load up the 3995 library with scratch and then copy the data from the ESS over to optical disks. Though it is also possible to restore from the ESS at the moment (once the data is copied to it).

Apparently, it is not taking as long as we had thought initially. It is now doing about 3 disks a day so not too bad.

I think the script should work with other media (such as tapes) because we had to give them details such as type of media (i.e optical) and type of interface (i.e SCSI or FC) and library details

 

[amalou]

he
i have the same library as you "3995" attached with scsi to AIX system the transfer speed are < 1Mbyt/s is it the same with you 3995