blackminder
ADSM.ORG Member
Good day to all!
I have not clear view about 'how to recover TSM from failure when something die
'
When my logmode = normal - i can use FULL/INC backups to restore (PIT restore), but if my logmode is rollforward ?
Main questions (note for all cases: DBBACKUPs (FULL/INC) always avail.):
case A:
If SERVER die (not disks), my action:
1) install new OS
2) copy original recovery log volumes to some location
3) install TSM, init TSM instance, shutdown TSM instance
4) useing DSMSERV loadformat to recreate dsmserv.dsk with same db and recovery log volume names and at the same location (as it was at old server, dsmserv.dsk)
5) copy devconfig, volhist and dsmserv.opt files to the TSM instace directory
6) OVERWRITE NEW created recovery log volumes with the original volumes. SIZE MUST BE EQUAL
7) run dsmserv restoredb
TSM restore db and roll transactions from RecoveryLog
or i can do steps 1,2,3 and overwrite new created dsmserv.dsk, db and log vols with old (copyed at step 2 + copy DB VOLs ) ?
Case B:
SERVER die, and one of mirrored copy for DB\LOG die too (one disk is die too)
at this case i cant copy back all recovery log volumes, because its die with disk. how can i restore TSM to the most current state ?
if i recreate DB and log VOLs, i cant copy all of RL vols.
Case C:
All die SERVER and disks, but i do periodical copy (just copy in widnows or cp in unix) for RL and DB vols. I think in this case restore as in case A, but with data loss because some transactions missed between COPY procedures. But its much better than nothing
i missed something ? where i wrong ?
I have not clear view about 'how to recover TSM from failure when something die
'When my logmode = normal - i can use FULL/INC backups to restore (PIT restore), but if my logmode is rollforward ?
Main questions (note for all cases: DBBACKUPs (FULL/INC) always avail.):
case A:
If SERVER die (not disks), my action:
1) install new OS
2) copy original recovery log volumes to some location
3) install TSM, init TSM instance, shutdown TSM instance
4) useing DSMSERV loadformat to recreate dsmserv.dsk with same db and recovery log volume names and at the same location (as it was at old server, dsmserv.dsk)
5) copy devconfig, volhist and dsmserv.opt files to the TSM instace directory
6) OVERWRITE NEW created recovery log volumes with the original volumes. SIZE MUST BE EQUAL
7) run dsmserv restoredb
TSM restore db and roll transactions from RecoveryLog
or i can do steps 1,2,3 and overwrite new created dsmserv.dsk, db and log vols with old (copyed at step 2 + copy DB VOLs ) ?
Case B:
SERVER die, and one of mirrored copy for DB\LOG die too (one disk is die too)
at this case i cant copy back all recovery log volumes, because its die with disk. how can i restore TSM to the most current state ?
if i recreate DB and log VOLs, i cant copy all of RL vols.
Case C:
All die
SERVER and disks, but i do periodical copy (just copy in widnows or cp in unix) for RL and DB vols. I think in this case restore as in case A, but with data loss because some transactions missed between COPY procedures. But its much better than nothing i missed something ? where i wrong ?
