Hi,
there is a bug when you install a cluster service sched , you have to open the
GUI and reenter the password.
When we install a new cluster schedule we map the aktiv node cluster disk to
the inactiv node using the same drive letter.
This way we can use the same opt and path to the opt to install the schedule on
the inactive Node.
We can also reenter the password on the inactive nodes' registry to make sure
the sched won't fail when the cluster moves resources
Mit freundlichen Grüßen / Best Regards
Markus Veit
_________________________________________
ScaleOn GmbH & Co. KG
SO-HS2-SBB
Leverkusen, B151 R.401
Tel.: +49 214 30 71254
Fax: +49 214 30 9671254
E-Mail: markus.veit AT scaleon DOT de
Internet : http://www.bayer-ag.de
A Bayer Company
An: ADSM-L AT VM.MARIST DOT EDU
Kopie:
Thema: AW: Problem with TSM
Client in a WIN2000 Cluster Environment
owner-adsm-l AT VM.MARIST DOT EDU
Received : 30.06.2004
16:12
Bitte antworten an "ADSM:
Dist Stor Manager"
-----Ursprüngliche Nachricht-----
Von: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] Im Auftrag
von Martin
Krauß
Gesendet: Mittwoch, 30. Juni 2004 15:35
An: ADSM-L AT VM.MARIST DOT EDU
Betreff: Problem with TSM Client in a WIN2000 Cluster Environment
Hello everyone,
We have some problems with our TSM scheduler service in a Windows 2000
Cluster Environment.
There are two cluster nodes in the cluster, and from each cluster node
we try to backup our data from the corresponding cluster-hard disks to a
remote TSM-server.
So cluster node A backs up data from the corresponding clustered
(failover) HD AA to the TSM node AAA and cluster node B who backs up
data from the corresponding clustered (failover) HD BB to the TSM node BBB.
We would like both these clusters to have failover-capabilities with
respect to each other, so that the backup from cluster node A
automatically switches to server 2 if it fails on server 1 and vice
versa with cluster node B
Our problem: two TSM scheduler services on each of the two servers
running the cluster are installed: One for cluster node A and one for
cluster node B. For some reason that I do not understand only the
second service to be installed installed works. The first service
installed (scheduler service for cluster node A on Server 1 and the
service for cluster node B on server 2) exits with an error (MS Event ID
4099 / Source: AdsmClientService / "Scheduler exited with a result code
of 12"). So the failover-mechanism on the cluster switches over to the
other node. There the second installed (working) service allows the data
to be backed up successully. But since the Schedule Service does not
work on the "original" nodes, the failover-capabilities of the cluster
are lost. We cannot switch cluster nodes between servers, because with
the failure of the schedule services installed first the cluster fails
and switches back to the nodes where the services that were installed
later are located producing the error decribed before.
The administrator of the TSM server tells me that the service seems to
transmit a wrong password - which cannot be, because I reinstalled it
with very simly default passwords and doublespellchecked the passwords.
Since he is a Linux expert, he has no experience with windows
clustering, and therefore cannot help me on this issue.
Does anyone have experience with such a clustered environment under
windows and know how to resolve it?
Thank you for your help,
Martin
--
********************************************
Martin Krauß
IT-Gruppe Geisteswissenschaften
LMU München
Ludwigsstraße 28
80539 München
Hotline: 2180 6400
Tel.:089 2180 1395
FAX: 089 2180 13543
Email: martin.krauss AT itg.uni-muenchen DOT de
********************************************
|
