Cluster Search Times Out (Bad Xml Error); A Device Name Is Required To Create A Cluster Partition; Cluster Resource Goes Comatose Immediately After Migration Or Failover; Smdr.novell Is Not Registered With Slp For A New Cluster Resource - Novell OPEN ENTERPRISE SERVER 2 SP 2 - CLUSTER SERVICES 1.8.7 FOR LINUX Manual

13.3 Cluster Search Times Out (Bad XML Error)

If you are using in Novell eDirectory 8.7.3x, time-outs are possible when you search from
TM
iManager for eDirectory objects (such as NCP Server objects, Volume objects, and Cluster objects)
because the Object Class attribute is not indexed by default. The LDAP sub-tree search can take
over 30 seconds, which causes the query to time out. For example, a Cluster object search from the
Cluster Options page returns the error:
Bad XML found during parsing when accessing cluster options
We recommend that you create a value index on the objects' Object Class attribute. (Object Class is
considered an attribute for indexing purposes.) This helps to reduce the time needed for the sub-tree
search from over 30 seconds to 10 to 50 milliseconds. For instructions, see "Creating an Index" in
the Novell eDirectory 8.8 Administration Guide.
Building indexes speeds up the sub-tree search, even if some partitions being searched do not
contain these types of objects. For example, searching for a Cluster object in a context that contains
only users is not expected to return results; however, the Object Class search is still performed, and
benefits from having an index present.
The sub-tree search performance issue is resolved in the eDirectory 8.8.x release with the addition of
the AncestorID feature.
13.4 A Device Name Is Required to Create a
Cluster Partition
If you are planning to work with shared-disk NSS pools and volumes, you must install a shared-disk
cluster by entering a device name for the cluster Split Brain Detector (SBD) partition at cluster
creation (new cluster) time. If you don't enter a device name, you won't be able to cluster-enable
NSS pools. On Linux, names are case sensitive.
13.5 Cluster Resource Goes Comatose
Immediately After Migration or Failover
When the SLP daemon ( ) is not installed and running on a cluster node, any cluster resource
slpd
that contains the command goes comatose when it is migrated or failed over to the
ncpcon bind
node because the bind cannot be executed without SLP.
For information, see "SLP" on page 33.
13.6 smdr.novell Is Not Registered with SLP for a
New Cluster Resource
You might get an error after creating a cluster resource indicating that is not
smdr.novell
registered with SLP for cluster resources, but the service for the node is registered.
smdr.novell
Error: "cluster--<212>: Read ResVol error -603"
Troubleshooting Novell Cluster Services 197