Yes, we just saw this last week. We had rebuilt an NT client (let's call
it XYZ) that lost its C: drive, but the restore had terrible throughput
(that's another topic). So, after it was all done, we wanted to test the
theory that maybe the NT client machine was causing the slow throughput.
We restored the same data to another NT machine (let's call it ABC), and
the restore went much faster, so we were happy that it wasn't a TSM
problem. However, this second restore did not complete 100%. Later that
night, the daily backup ran for XYZ, but failed because "a recoverable
restore for this node is pending". SO, I think that's the answer... the
server keeps track of recoverable restores, and won't let a backup run at
the same time. Now, I don't know what happens for a NON-recoverable
restore...
Also, come to think of it, we did the restore on machine ABC by using the "
-nodename XYZ" parameter... I wonder if the same check would occur if we
used the "access another node" from the menu?
used the "access another node" from the menu?
Robin Sharpe
Berlex Laboratories
"Martin, Jon
R."
<jrmartin@KNS To: ADSM-L AT VM.MARIST DOT EDU
.COM> cc: (bcc: Robin Sharpe/WA/USR/SHG)
Subject:
07/17/01 Re: What would you think would
happen..........
03:00 PM
Please
respond to
"ADSM: Dist
Stor Manager"
Zoltan,
Interesting subject I was just talking about this possibility the other
day.
My question refers to part 1. When you say the node is not a TSM-client I
am assuming that they have the TSM client installed they are just not
registered with the server as a node? Otherwise how would they access the
utilities to open the connection in the first place?
Besides taking the safety precaution of locking the source node out has
anyone experienced this situation before?
Jon
|
