Odd Result when re-defining Library

[etchingsj]

Ok So here is the third time I have seen this in my TSM environment.

Some background:

TSM 5.4.4 On Windows 2003 Server
500+ nodes
2 TS3310 Libraries
2 TSM servers.

When we added drives or have made significant changes to our TSM environment and rebooted libraries, fiber switches or the Windows TSM server some times drives go offline due to the change of path in Windows.
What was mt1.0.1.2 is now mt0.0.1.2. As you can imagine this is an issue as you will get IO errors in TSM.


So IBM is very familiar with this error and their advice is the following.

export the following information

q drive f=d
q path f=d

Delete ALL paths
Delete ALL drives
Delete library path
Delete library

Run tsmdlst on the windows server.

Define the library
Define the libary path
Define ALL drives
Define ALL drive paths

I understand this operation but the next step is what is interesting.
One IBM support person told me to audit my library next, which tests the ability to use your drives and this works.

But on another occasion I was told to run manual checkin's for all my scratch and private volumes.

When I started this process I know I had 5 scratch tapes but when this process ends with the "checkin" and not the "audit" I always find 20 to 30 scratch tapes????

Where did they all come from?
So could I just do this once a month and "find" 30 tapes??? Did these tapes become mismarked as Private?
Maybe I should just check out everything in the library one a month and then run a checkin?

Has anyone had a similiar experience?

Thanks in advance,




​

 

[toxy13]

if I unterstand you: you have 5 tapes in your library, you run a checkin command and the result is: TSM had checked in 20-30 tapes?

 

[etchingsj]

No, not quite.

When the process began my library was near capacity with 211 tapes, but only 5 were marked as 'Scratch'. After the changes to re-path and redefine steps I ran a manual checkin. This is when the 30 scratch tapes appeared.

So somehow 25 tapes magically appeared.

I plan to look at my actlog and volhist to determine what happened with these volumes.

But this is the third time this has happened.

 

[toxy13]

Stupid question: you are sure, that nobody put 25 tapes in your library?

 

[etchingsj]

When I first inherited TSM and these libraries that was exactly my thought, but now 1 year later Nobody in our current IT team touches TSM but myself.

So, it is a very valid question but that is not the issue this time around.

 

[toxy13]

No, not quite.

When the process began my library was near capacity with 211 tapes, but only 5 were marked as 'Scratch'. After the changes to re-path and redefine steps I ran a manual checkin. This is when the 30 scratch tapes appeared.

So somehow 25 tapes magically appeared.

I plan to look at my actlog and volhist to determine what happened with these volumes.

But this is the third time this has happened.

could it be, that tape with status "pending" will go to scratch by a new checkin? (reuse delay)

 

[etchingsj]

Upon further investigation I have noticed when the Windows TSM server was restarted again the extra scratch tapes vanished.

In searching my activity log I realized they have been marked "read only". I went through the process of checking all those volumes out and then running a label with the search command and they once again marked "scratch"

At this point I will just monitor to see what becomes of this media over the next week.

Thanks