Server Processes Waiting for Mountpoints

[Silverbell]

Hi,
This is such a silly question but i can't find the answer anywhere... at least using my search terms.

What server setting, device class setting, or storage pool setting, etc... Controls HOW long a Server Process will continue to wait for a mount point before timing out? I am referring to Migration, Reclamation, DB Backup, backup stgpool etc... I need to control how long these processes will continue to wait for a mountpoint .

From what i am reading - the mountwait parameter only has to do with fulfilling mount requests and NOT to do with my question above...

I am sure this is a simple answer but ill be darned if i can find it.

Thanks

 

[heada]

A server process will wait forever unless there is a time limit placed on it (like the duration option for migrations) or the server thinks that the server is out of storage.

Client processes (backup, restore, etc) will use the mountwait setting.

-Aaron

 

[Silverbell]

Thank you as always Aaron,
Would this Server Processing unlimited WAIT condition apply to the BACKUP NODE command as well? Being Run from the Server via an ADMIN Schedule where client writes directly to tape via datamover...

Best ~~ Silverbell

 

[heada]

I've never tried it and I honestly don't know. Normally it would be a client function (even if the server prompted it) but since you're going over a datamover, it might be considered a server function.

Try it out and let us all know.

-Aaron

 

[reddy]

Thank you as always Aaron,
Would this Server Processing unlimited WAIT condition apply to the BACKUP NODE command as well? Being Run from the Server via an ADMIN Schedule where client writes directly to tape via datamover...

Best ~~ Silverbell

I like to have something like that for BACKUP NODE but did not find anywhere. If you are in 5.3.4.x or 5.4.0 try upgrading to 5.4.1...from this version ndmp timeouts changed from 10mins to 3hrs for a mount request. Nowhere this 'time out' can be set except in the tsm code which can only be done by IBM/tsm developers...

Hope this helps if I understand your question correct.

 

[Silverbell]

We are on 5.5 COde so i guess the 3-hour limit applies?

Where did that info come from? That was exactly what i was looking for!!

Thanks to all of you - I love adsm.org

 

[reddy]

Hope 3-hr parameter covered in 5.5 as well...but not sure about that. Our tsm servers are at 5.4.1 on AIX platform. Make sure the 'mountwait' is set to equal or more than 180min. Or else you could do one thing...define a tsm script to backup fewer filespaces at a time. For example...

line 1) backup node FILER1 filespace1,filespace2,filespace3 toc=yes wait=yes
line 2) backup node FILER1 filespace4,filespace5,filespace6 toc=yes wait=yes


Include filespaces in each line depending on how many drives it can use at a time.

Or you can check with IBM if that 3-hr time out is applicable in version 5.5