Restoring Database to new HW Config

Thank you all so much!
I think I'm ready for a hands-on crashcourse in TSM DR now.
Wish me luck.

Come back and post details/experience on how it goes, good/bad/ugly and what could be done better to improve, etc.
 
Come back and post details/experience on how it goes, good/bad/ugly and what could be done better to improve, etc.


Certainly will do! I've got to create a doc from scratch w/ screenshots, etc for this exercise, so I'll have plenty of details engrained in my head. I oughta be able to give a pretty good recap and some tips.

Do you happen to know the answer to this:
Which file lists the volumes needed, is it XXXXX.COPYSTGPOOL.VOLUMES.AVAILABLE.MAC? When I break up the drplan file there's about 25 different files.
 
Certainly will do! I've got to create a doc from scratch w/ screenshots, etc for this exercise, so I'll have plenty of details engrained in my head. I oughta be able to give a pretty good recap and some tips.

Do you happen to know the answer to this:
Which file lists the volumes needed, is it XXXXX.COPYSTGPOOL.VOLUMES.AVAILABLE.MAC? When I break up the drplan file there's about 25 different files.

Volumes for what? The volhist and dbb will have all the info needed. I believe that list provides a list of whats in the copystgpool at the time the prepare was run, however I could be wrong. I've never used those and have always gone off the volhist/devconfig as well as the dbb for restore/DR purposes.
 
When I break up the drplan file there's about 25 different files.

Chapter 24 is great.
http://publib.boulder.ibm.com/infocenter/tivihelp/v1r1/topic/com.ibm.itsmcw.doc/anrwgd55789.htm#udrm

Specifically, toward the end of the chapter is info on the Planfile and an example and explanation of all the parts. You "could" run the entire plan in one command, but in DR testing, I'd recommend stepping thru it manually, watching each thing happen before moving to the next. That will also give you a hands on experience in what happens during the recovery script. I did my first DR test 2 years ago and had never done it, DRM or otherwise. Dug into that documentation to build our plan and the DR test worked great.
 
Last edited:
Volumes for what? The volhist and dbb will have all the info needed. I believe that list provides a list of whats in the copystgpool at the time the prepare was run, however I could be wrong. I've never used those and have always gone off the volhist/devconfig as well as the dbb for restore/DR purposes.

THe volumes I'm talking about are the ones that I send to the DR Vault daily. I guess they contain data to restore client machines.


Chapter 24 is great.
http://publib.boulder.ibm.com/infoce...55789.htm#udrm

Specifically, toward the end of the chapter is info on the Planfile and an example and explanation of all the parts. You "could" run the entire plan in one command, but in DR testing, I'd recommend stepping thru it manually, watching each thing happen before moving to the next. That will also give you a hands on experience in what happens during the recovery script. I did my first DR test 2 years ago and had never done it, DRM or otherwise. Dug into that documentation to build our plan and the DR test worked great.


Thanks for the link to the guide. I'll check that out as well. I think this will go at least half decently..
 
Ok, I'm ready for more help.
Do I need the volhist.out, etc files in place on the new server BEFORE I do a DB restore???

I ran the command "c:\program files\tivoli\tsm\server\dsmserv" restore db devclass=dbb-pit vol=79535912.dbb commit=yes

and I got
The ICC Support module has been loaded.
Unable to read disk definition file dsmserv.dsk
HBAAPI vendor library failed to load or is missing
 
Ok, I'm ready for more help.
Do I need the volhist.out, etc files in place on the new server BEFORE I do a DB restore???

I ran the command "c:\program files\tivoli\tsm\server\dsmserv" restore db devclass=dbb-pit vol=79535912.dbb commit=yes

and I got
The ICC Support module has been loaded.
Unable to read disk definition file dsmserv.dsk
HBAAPI vendor library failed to load or is missing

Attempt to run the command in "c:\program files\tivoli\tsm\server\" directory, i.e. cd into \program files\tivoli\tsm\server\ and run and see what happen.
 
Attempt to run the command in "c:\program files\tivoli\tsm\server\" directory, i.e. cd into \program files\tivoli\tsm\server\ and run and see what happen.


Ok, gave that a shot and this was returned:
ANR47261 The ICC support module has been loaded
ANR8208W TCP/IP driver unable to initialize due to error in using port 1500 reason code 10048
ANR0212E Unable to read disk definition file dsmserv.dsk
 
Funnily enough, I was doing a Windows DR test for a customer only last week so some of this is quite fresh. I'm jumping in right at the end of the thread without reading the pages previous, so apologies if any of this has been covered.

Ok, so you might want to be running this from the directory in which your config files are - on Windows, this might be in Program Files\Tivoli\tsm\Server1 or Server2 directories, not the 'server' directory where the binaries live. So your command might look like ..\server\dsmserv restore todate=<etc>.

The "ANR8208W TCP/IP driver unable to initialize due to error in using port 1500 reason code 10048" suggests you might already have something else listening on port 1500. Have you another TSM Server running on this server already showing in 'services' or 'net start'?

If you do have more than one TSM Server instance then you'll need to invoke the dsmserv command with a '-k <servername>' option to point dsmserv at the correct registry key for your TSM Server instance (dsmserv on its own defaults to -k server1) and be in the corresponding Server1/Server2 directory when you run dsmserv.

Finally, have you pre-defined any database or log volumes into which to perform your recovery?

Cheers,

//David Mc
London
 
Sorry it took so long to reply to this, it's been a rough week so far.
Anyways, the TSM Server1 service is running from C:\Program files\tivoli\tsm\Server\dsmsvc.exe so that's where ran DSMServ from. I'm not even sure who set this site up, but I may be redoing it.
Also I did make sure to CD to the directory where dsmserv resides before running the restore command.

Edited to say, that I tried running the restore command using dsmsvc, not sure if during installation you can rename that, but it accepted the command, but there was no output in the CLI and nothing showed up when issuing a q pr in the server CLI.
Also in the TSM Management console it says that TSM Server1 is installed to c:\program files\tivoli\tsm\server..

Maybe I just need to kill this thing and start from scratch.

Funnily enough, I was doing a Windows DR test for a customer only last week so some of this is quite fresh. I'm jumping in right at the end of the thread without reading the pages previous, so apologies if any of this has been covered.

Ok, so you might want to be running this from the directory in which your config files are - on Windows, this might be in Program Files\Tivoli\tsm\Server1 or Server2 directories, not the 'server' directory where the binaries live. So your command might look like ..\server\dsmserv restore todate=<etc>.

The "ANR8208W TCP/IP driver unable to initialize due to error in using port 1500 reason code 10048" suggests you might already have something else listening on port 1500. Have you another TSM Server running on this server already showing in 'services' or 'net start'?

If you do have more than one TSM Server instance then you'll need to invoke the dsmserv command with a '-k <servername>' option to point dsmserv at the correct registry key for your TSM Server instance (dsmserv on its own defaults to -k server1) and be in the corresponding Server1/Server2 directory when you run dsmserv.

Finally, have you pre-defined any database or log volumes into which to perform your recovery?

Cheers,

//David Mc
London
 
Last edited:
Hi,

seems you do not have your DB/RLOG volumes formated and therefore dsmserv.dsk (descriptor file containing the list of DB and RLOG volumes with a hash) is missing.
Two ways
1) create the volumes using "dsmserv format" and dsmfmt
2) create new - blank - TSM instance, create new DB and RLOG volumes using normal TSM commands, stop the instance, replace devconfig.txt and volhist.out with original ones and restore the DB.

I prefer the second one ...

Harry
 
Last edited:
Hi,

seems you do not have your DB/RLOG volumes formated and therefore dsmserv.dsk (descriptor file containing the list of DB and RLOG volumes with a hash) is missing.
Two ways
1) create the volumes using "dsmserv format" and dsmfmt
2) create new - blank - TSM instance, create new DB and RLOG volumes using normal TSM commands, stop the instance, replace devconfig.txt and volhist.out with original ones and restore the DB.

I prefer the second one ...

Harry


I may give this a shot, so rather than reinstall everything, just create a new server instance. That's kind of my boss' plan anyway, to have the DR TSM Server be able to back up the other DR servers in case just one or two of them have to go online and our Primary TSM Server is still up, but still have the option of restoring our Primary to the DR in case of the Primary TSM Server completely failing.
 
There is no DSMSERV.EXE in the SERVER1 folder on either of my installations.. I've reinstalled the server at my DR Site and I don't even see any way to get the file installed to that directory, it always goes to the SERVER directory.
 
"There is no DSMSERV.EXE in the SERVER1 folder on either of my installations.. I've reinstalled the server at my DR Site and I don't even see any way to get the file installed to that directory, it always goes to the SERVER directory."

That's the point I was trying to make above - I don't believe there's supposed to be. The DSMSERV.EXE lives in the SERVER directory, and then for each new TSM Server instance a new Server1, Server2 etc directory gets created. If you want manually to perform actions from the CLI, it's best to be in the (e.g.) Server1 directory and then issue your DSMSERV.EXE commands by ..\SERVER\DSMSERV.EXE FORMAT | RESTORE | QUERY etc.

Take a look at this link - the bottom Task explains much more eloquently what I'm trying to explain:

http://publib.boulder.ibm.com/infoc.../com.ibm.itsmcw.doc/b_adminref_windows558.htm

Or have I got the wrong end of the stick with what you're trying to achieve?

HTH,

//David Mc
London, UK
 
I'm getting somewhere now.. I had to stop the services for Server1 and that got the recovery to start. I had the services running when trying to restore the DB before. Had to copy my devcnfg file too, because I reinitialized my server1. Also got a hang up because my initial DB file size used when I set up the new instance of Server1 was too small. I'm about to redo server1 again and make sure the initial DB size is large enough and cross my fingers. I think I'm just about there!

"There is no DSMSERV.EXE in the SERVER1 folder on either of my installations.. I've reinstalled the server at my DR Site and I don't even see any way to get the file installed to that directory, it always goes to the SERVER directory."

That's the point I was trying to make above - I don't believe there's supposed to be. The DSMSERV.EXE lives in the SERVER directory, and then for each new TSM Server instance a new Server1, Server2 etc directory gets created. If you want manually to perform actions from the CLI, it's best to be in the (e.g.) Server1 directory and then issue your DSMSERV.EXE commands by ..\SERVER\DSMSERV.EXE FORMAT | RESTORE | QUERY etc.

Take a look at this link - the bottom Task explains much more eloquently what I'm trying to explain:

http://publib.boulder.ibm.com/infoc.../com.ibm.itsmcw.doc/b_adminref_windows558.htm

Or have I got the wrong end of the stick with what you're trying to achieve?

HTH,

//David Mc
London, UK
 
When I put the devcnfg.out file into the server1 directory why does it gives me the devices needed to restore the DB, but none of the devices show up in the ISC?
 
When I put the devcnfg.out file into the server1 directory why does it gives me the devices needed to restore the DB, but none of the devices show up in the ISC?


Anyone know the answer to this one?

Also, I'm having some success just following the drplan file for disaster recovery mode. Still running into some snags though. I'll update later.
 
I can't comment on the ISC I'm afraid - rightly, or wrongly, I've completely avoided the ISC with all customers I've worked with in the 5 or 6 years it's been around. Certainly during a DR situation I'd prefer to drive things from the CLI anyhow, but that's just personal preference.

Sounds like you're nearly there... I was doing a Windows TSM DR test with another customer just this weekend - even after all of these years you still come across little snags in DRs that keep you scratching your head...

Good luck.
_________________
/David Mc
London, UK
 
Yeah.. Just finished the whole DB restore, and when the server attempted to start I got errors that the Library cannot be found. It's looking for my tape library at the main site, which isn't even listed in my defcnfg file. Not sure why? I've updated my devcnfg and my server options file. THis sucks.. It's right at the END and I keep running into this..
Also, during a DR will the server create disk volumes, etc that it needs to get itself back up and running?

WHat a heartbreaker. I'm going to try again after defining the library in the device configuration wizard. If it errors out again I can probably provide some more details of the messages I'm getting at the end.
 
When I put the devcnfg.out file into the server1 directory why does it gives me the devices needed to restore the DB, but none of the devices show up in the ISC?

So at this point was the restore of the DB/TSM Server successful? If so had you tried rebooting? The entry in the ISC may have been updated, or the instance removed and readded with the "updated" status.

We normally stay clear of the ICS as well...but theres a possiblity all devices (paths/drives) have to be remapped because the OS seems them different than the other OS did.
 
Back
Top