Disaster Recovery (Part 2)
1995-01-10 03:46:45
Hello again,
This post is closely related to the one of yesterday (Migration Threshold) and
expose one more problem I've encountered in doing our disaster recovery
procedure:
The procedure is the following :
1)Sunday morning we disable the server
2)We kill all active (if any) non-admin sessions
3)We start the migration to cartridge STG (see post on HIGHMIG limitation)
4)When done, we start a DUMP DB CONSISTENT=YES
5)At the same time we start to copy every tape of the Tape Stgs to a set named
"the outgoing security set".
6)When done we re-enable the server
The problem is in step 5 : Apparently ADSM writes data to Cartridges up to
the physical E-O-T (End of Tape) regardless of the ESTCAP parameter I gave in
the definition of DEVICE CLASS (where I did specify (real_capacity)*0.95 to
cope with the possible problem of different cartridge length)
In this way I'm not sure that my input tape (Status FULL in Q VOL) will fit
onto the output tape ..
I already encountered this copy problem twice and the set of tapes is hence
unusable in case of server reconstruction.
The question everybody is going to ask me is : why are u not using EXPORT
SERVER ...
I think that previous posts on the duration of the EXPORT SERVER can explain
why ...
Cheers, michel
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- Disaster Recovery (Part 2),
Michel Gaillard <=
|
|
|