virtual volume puzzler

[kimon]

I've got two backup servers, once source, one target. I have a storage pool on the source server that I want to copy to the target server.



When I run the command:



backup stgpool SP.SIG.DISK SP.SIGCOPY.EXTERNAL wait=yes



I get the following in actlog:



10/23/2002 17:23:56 ANR2110I BACKUP STGPOOL started as process 83.

10/23/2002 17:23:56 ANR1210I Backup of primary storage pool SP.SIG.DISK to

copy storage pool SP.SIGCOPY.EXTERNAL started as process

83.

10/23/2002 17:23:59 ANR4370E The source server is not authorized to create

volumes on target server SERVER.SIGCOPY.

10/23/2002 17:23:59 ANR1404W Scratch volume mount request denied - mount

failed.

10/23/2002 17:24:01 ANR4370E The source server is not authorized to create

volumes on target server SERVER.SIGCOPY.

10/23/2002 17:24:01 ANR1404W Scratch volume mount request denied - mount

failed.

10/23/2002 17:24:03 ANR4370E The source server is not authorized to create

volumes on target server SERVER.SIGCOPY.

10/23/2002 17:24:03 ANR1404W Scratch volume mount request denied - mount

failed.

10/23/2002 17:24:03 ANR1216E BACKUP STGPOOL: Process 83 terminated - storage

media inaccessible.

10/23/2002 17:24:03 ANR0985I Process 83 for BACKUP STORAGE POOL running in the

FOREGROUND completed with completion state FAILURE at

17:24:03.



Two questions:



1. Why does it need to create a new volume on the target server? I've defined a new volume on the target server, its empty, and hooked up to the correct storage pool. BTW, the storage pool on the target server is a DISK pool. Also, the associated copygroup is of type=archive.



2. If it indeed does need to create new volumes, how do I giive the source server authority to do so?



For those interested in the gory details, here is my setup:



SOURCE SERVER SETUP:



define server SERVER.SIGCOPY

password = nodepasswd

hladdress = 10.2.1.165

lladdress = 1500

nodename = SNODE.SIGCOPY

define devclass DC.SIGCOPY.EXTERNAL

devtype = SERVER

servername = SERVER.SIGCOPY

define stgpool SP.SIGCOPY.EXTERNAL DC.SIGCOPY.EXTERNAL

pooltype = copy

maxscratch = 1

define stgpool SP.SIG.DISK DISK

cache = yes

copystgpools= SP.SIGCOPY.EXTERNAL

define volume SP.SIG.DISK C:\tsmdata\vol.sig.dsm

formatsize = 1

wait = yes





TARGET SERVER SETUO:



define stgpool SP.SIGCOPY DISK

define volume SP.SIGCOPY C:\tsmdata\vol.sigcopy.worm.dsm

formatsize = 2

wait = yes



define domain DOM.SIGCOPY

define policyset DOM.SIGCOPY POL.SIGCOPY

define mgmtclass DOM.SIGCOPY POL.SIGCOPY MC.SIGCOPY

spacemgtech = auto

migrequiresbkup = no

assign defmgmtclass DOM.SIGCOPY POL.SIGCOPY MC.SIGCOPY

define copygroup DOM.SIGCOPY POL.SIGCOPY MC.SIGCOPY

type = archive

dest = SP.SIGCOPY

activate policyset DOM.SIGCOPY POL.SIGCOPY



register node SNODE.SIGCOPY nodepasswd

domain = DOM.SIGCOPY

passexp = 0

type = server





Thanks for any ideas...



Kimon

 

[run]

yes, the server to server communication was always a bit of a bitch to setup. Always had problem, when I was teaching TSM class with students messing this up (because in my opinion it was too complicated).



Anyway, here's some hints that will help you along.



open a "dsmadmc -console" (console) window on both servers and watch the consoles

for informative error messges.



Remember that the server password is distinct from the nodename password. That doesn't mean you can't set both to the same thing. (see help define server for more details). Also, I also didn't see that "set" command (though you could have run them but forgot to put them in the post)



(of course, this is just an example from the help command, do this for each server)

set servername server_b

set serverpassword mylife

set serverhladdress 9.115.20.80

set serverlladdress 1860



Lastly, start off a bit more basic in your testing. Try doing a database backup to the server device class and then once you get this working, then add a bit more complexity and try using a storage pool based on that working device class.



Yes, it's complicated and remember that the virtual volume as it is referred to on the remote server is seen as one large archive file. Meaning you can't reclaim it unless you send the whole archive file back to the original server which then unpacks that archive file into it's individual files. Reclaims the space and sends the reclaimed/compacted files back to the original server. (yes, your network admin will love you for doing this But, if you only send your database backups and other special stuff instead of using it for a main copy storage pool, then I can see a good benefit in that.



Otherwise, good luck and have paitence.... "use the force luck" but don't force it sometimes it only took me an hour to get it workin in class...though some students gave up after 2 hours ..



rick

 

[kimon]

Rick, thanks for the tip about -console, that is much easier that looking at the actlog.



> Remember that the server password is distinct from the nodename password. That doesn't

> mean you can't set both to the same thing. (see help define server for more details). Also, I

> also didn't see that "set" command (though you could have run them but forgot to put them in the post)



Yea, you know I actually did have the password wrong before I posted, but I had seen the access error in the log file and fixed it, so I know that is no longer the problem. Also, just to be sure I am now using the same password for nodes and servers.



> (of course, this is just an example from the help command, do this for each server)

> set servername server_b

> set serverpassword mylife

> set serverhladdress 9.115.20.80

> set serverlladdress 1860



Yes, these I did.



> Lastly, start off a bit more basic in your testing. Try doing a database backup to the server

> device class and then once you get this working, then add a bit more complexity and try using

> a storage pool based on that working device class.



I did as you succested and tried the following command:



backup db type=full devclass=DC.SIGCOPY.EXTERNAL



The errors in the seem identical:



ANR2280I Full database backup started as process 111.

ANR4370E The source server is not authorized to create volumes on target server

SERVER.SIGCOPY.

ANR1404W Scratch volume mount request denied - mount failed.

ANR4578E Database backup/restore terminated - required volume was not mounted.

ANR0985I Process 111 for DATABASE BACKUP running in the BACKGROUND completed

with completion state FAILURE at 13:27:31.



Any other ideas as to why it things it needs to create a volume when one already exists? Can it be that my destination pool on the target server is type DISK, which it not what it expects because it is supposed to be a copy pool?



BTW, I tried the test again, but this time created three volumes on the target server of 100MB each, and still the same problem... :cry:



Kimon

 

[run]

ok, I'm trying to remember but I get a vague feeling that the archive copy group on the destination server must point to sequential media (tape or sequential file).



The only way that I got it to work in class was using a storage pool of dev type = file (sequentail file) at the destination server. I've never tried (or got it work...I think) using a disk storage pool at the destination side.



Give that a go and see what happens.

 

[kimon]

<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font class="pn-sub">Quote:</font><HR></TD></TR><TR><TD><FONT class="pn-sub"><BLOCKQUOTE>ok, I'm trying to remember but I get a vague feeling that the archive copy group on the destination server must point to sequential media (tape or sequential file).



The only way that I got it to work in class was using a storage pool of dev type = file (sequentail file) at the destination server. I've never tried (or got it work...I think) using a disk storage pool at the destination side.



Give that a go and see what happens.

</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR></TD></TR></TABLE>



Failes in the same way, unfortunatly.



ANR4370E The source server is not authorized to create volumes on target server

SERVER.SIGCOPY.



Looks like it is trying to create a scratch volume for the backup but is unable to.



Is it possible to pre-create "scratch" volumes for a storage pool with a file device class?

If so, perhaps that is what I am missing.



Kimon