| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 422.1 | | DECWET::ONO | Software doesn't break-it comes broken | Wed Feb 19 1997 08:47 | 18 |
| Gene,
Welcome to NetWorker Support. Bob told me you and Bruce had
joined up.
If the file system is not in /etc/fstab, then NetWorker won't
pick it up automatically using a saveset of "All". You might
have the customer try setting up a separate client resource with
the saveset set to the specific filesystem, and see if that will
pick it up. You may have to put this other client into another
savegroup, since the "All" from the first client may conflict
with the specific saveset in the second.
Real cluster support is a client-side function. I'll check with
Legato tech support to see if the current HP client understands
HP clusters. I think it doesn't.
Wes
|
| 422.2 | Legato incident 246339 | DECWET::ONO | Software doesn't break-it comes broken | Thu Feb 20 1997 08:24 | 0 |
| 422.3 | | DECWET::ONO | Software doesn't break-it comes broken | Thu Feb 20 1997 16:28 | 70 |
| I got a response from Legato tech support that asks a lot of
questions. Maybe you can review this in your configuration.
As far as the question of how ServiceGuard compares against ASE,
there's no way to tell. Also, since cluster support is almost
completely client-side, the server doesn't matter all that much.
Wes
Response from Legato tech support:
========================================================================
This is in response to the above logged INCIDENT NUMBER that you have open
with us. The incident has to deal with:
A question you had about the NetWorker support of "ServiceGuard" clustered
HP environment.
Recommendations:
----------------
We recommend/suggest the following:
Well as far as I can tell we have not tested "ServiceGuard" with Networker
but that does not mean that it does not work. As a matter clusters in an
HP environment have been known to work in the past. Here is an example of
a customer setup using HP clusters:
Env: NW 4.2 on an HP-UX 9.05 9000/755 server. The customer was using CDF's
for the /nsr file systems. The reason was that he had two networker servers on
the same cluster and his configuration then was:
machine BEAR - Networker server and a client on a HPUX cluster
/nsr+/bear/index
/nsr+/bear/cores
/nsr+/bear/jukebox
/nsr+/bear/logs
/nsr+/bear/mm
/nsr+/bear/res
/nsr+/bear/tmp
machine IROCZ - Networker server and a client on a HPUX cluster
/nsr+/irocz/index
/nsr+/irocz/cores
/nsr+/irocz/jukebox
/nsr+/irocz/logs
/nsr+/irocz/mm
/nsr+/irocz/res
/nsr+/irocz/tmp
Without CDFs of /nsr, he would have a conflict of information between the
machines. There were no problems backing up the hosts and even doing recoveries
on the cluster nodes. I need more information from the customer as to his
configuration set up and here it is listed:
1. OS releases of Networker Server and client. What version of Networker are
running on the server/clients ?
2. Need some information about his cluster configuration. How many hosts has
he configured in his cluster and how are they being backed up.
3. How are things being synchronized among the various cluster nodes during
the backups. I am not veri familiar with the "ServiceGuard" product and need
to ask the details from your side.
How does ServiceGuard product compare with DECsafe ASE/TruCluster ... ? Does
Dec's version of NetWorker support this product ?
For your reference, the INCIDENT NUMBER for this matter is 246339.
Thank you for your cooperation to get this incident resolved.
|
| 422.4 | | DECWET::ONO | Software doesn't break-it comes broken | Fri Feb 21 1997 12:19 | 3 |
| See workaround in 431.2
Wes
|