cancel
Showing results for 
Search instead for 
Did you mean: 

netware cluster support

clod
Level 3

you can't back up a netware cluster effectively with be12. one failover since

selection list was made and the job's shot, not to mention the mess when

you need to recover.

why is there no support for this ? symantec/veritas have known how to handle this

for a long time.

 

5 REPLIES 5

clod
Level 3

no response, what a shocker.

 

I saw several people/posts hinting that this support was "just around the corner" (pp)

and may appear in 12.5.

wishful thinking.

 

could someone from symantec just let me know whether this is actually going to happen

and if so, when ?

 

need to know whether we should drop BE or not.

it can't backup our system, so as backup software, it's not what I would

call good value for money right now.

 

quite the reverse in fact.

 

many thanks.

 

 

 

 

 

clod
Level 3

I've seen messages in other threads stating that symantec watch these forums

and they're keen to understand the shortcomings of their software and how it

is used.

 

I've given my feedback and asked whether or not BE for windows servers will

ever gain support for netware (and by extension, OES2/linux) clustering.

 

I don't appear to be able to get an answer from anywhere else,

so how about throwing me a bone ?

I feel I've paid enough money to you over the years to deserve one.

 

 

 

 

 

 

 

 

Turls
Level 6

Clod!

 

I have worked with the support team to get a response...

 

"Thereisn't much we can from a technical support perspective.  The product is working as designed.  Since going to an NDMP agent for NetWare withversion 10.0, BEWS is not able to backup NetWare cluster volumes as clusterresources.  It is backed up as a localvolume from the active node.  If thevolume moves to a different node, the job will fail. 

 

The only workaround to make sure you get a backup is to setup a job with allcluster volumes on one node.  Then setupanother job with all cluster volumes on the other node.  When the jobs run each night, one will alwaysfail because it is not the active node for the volume.  If a failover occurs without a backuprunning, it will backup when the newly active node runs.  If a failover occurs during the backup, youare out of luck."

Since this is anenhancement request, we shall work with the PM regarding this.

 

Hope this helps.

 

Swathi

Community Manager

Backup and Archiving

Message Edited by Turlas on 01-28-2009 11:13 AM

clod
Level 3

hello Turlas,

 

thanks for following this up.

following the suggested workaround means that with a typical 5 node cluster here,

we'd have every job fail as soon as a single resource is moved,

5 failed jobs logged daily, which would of course make it very difficult (extremely) to spot genuine failures.

incremental backup jobs wouldn't work, and as a full backup on this config

takes over a day, it means a daily backup isn't practical.

locating a file for recovery from tape in amongst the failures wouldn't be much fun to say the least.

I'd really say that although BEWS 10 may be working as designed, without making a few changes,

it's unsuitable for netware/suse clusters. 

 

failovers occuring during backup really isn't an issue, even if it does fail that resource for that job.

the issue is having BEWS locate the volumes carrying the cluster resource to be backed up

when the job is started (from the output of cluster status or nudp display even) and then marrying

the cluster resource up to previous backups so incremental backups can work.

it doesn't seem like a huge amount of changes to do this, ndmp transport doesn't really come into it,

it sounds like it can all be done in the client.

 

 

many thanks

 

clod
Level 3
 is there any update to this ?
as be for netware has been dropped there's just no usable way of backing up the typical netware installation with BE now (netware and oes2/linux clusters pretty much being the norm)
as BENW is no more, this is pretty urgent.
it's been nearly 3 months since you said you were working on it, surely that's long enough
to have a rough idea of what's happening.