On Mon, 27 Aug 2007 22:04:54 +0000, Ben Knorr wrote:

> I've been tinkering with the cluster for a little while, and have learned
> a few things.
> I set up an iprint driver store and iprint manager on printcl1. I
> manually ran through each line in the unload script in the iprint resource
> for the cluster, and as far as I can tell, everything went fine. I then
> stopped cluster services on printcl1 and then checked to see if the iprint
> driver store/manager switched to printcl2.
> The driver store states that it is down and that it did not reach the
> desired state within the alloted time. I checked
> /var/opt/novell/log/iprint/idsd.log ,and see a fatal error regarding idsd:
> Aug 27 15:31:01 INFO IDSTerm: TermRepositoryInfoList() completed Aug 27
> 15:31:01 INFO IDSTerm: DeinitIdsJobQueue() completed Aug 27 15:31:01 INFO
> IDSTerm: DeinitIPC() completed Aug 27 15:32:04 INFO Building Driver Table
> of Contents. Aug 27 15:32:04 INFO Driver Table of Contents Complete. Aug
> 27 15:32:04 INFO idsd operational - listening for connections... Aug 27
> 15:32:57 FATAL Another instance of idsd is already running [1137]

You should make sure that iPrint is not loading on the box during boot
(which is the default). Try

Offline your resource
rcnovell-ipsmd stop
rcnovell-idsd stop
chkconfig novell-ipsmd off
chkconfig novell-idsd off

> I've learned a bit about how the cluster works (apparently through links
> in /var/opt/novell/ to the shared volume),

Yep, well spotted - you don;t want to hear my thoughts on that :-O

> but still am perplexed about
> my various .conf files in /mnt/iprint/etc/opt/novell/iprint/conf
> directory (idsd.conf and printmanager.Printing.wcslc_ut:
> ipsmd.conf:DSServer1 printcl1.westminstercollege.edu ....
> Does the ncs_iprint_svc_start change this conf file to reflect which
> server is hosting the manager/driver store? If not, how is this
> supposed to be set up so that the proper server is listed in these
> files? In this case, when the manager/driver store is rolled over to
> printcl2 when printcl1 goes down, at which point does printcl2 figure
> out it is where these resources (which apparently are stored on the
> shared volume) are located?

The DS server is the ldap server that you specified during the creation of
the resource. I always use a non-cluster node for that. You should be
able to manualy edit this file and point it elsewhere.

Mark Robinson
Novell Volunteer SysOp
One by one the penguins steal my sanity...