IBM SG24-5131-00 Manual Do Utilizador
174
IBM Certification Study Guide AIX HACMP
Be aware that persistent sticky location markers are saved and restored in
cluster snapshots. You can use the
cluster snapshots. You can use the
clfindres
command to find out if sticky
markers are present in a resource group.
If you want to remove sticky location markers while the cluster is down, the
default
keyword is not a valid method, since it implies activating the resource.
Instead, when the cluster is down, you use a transient
stop
request, as in this
example:
(The optional -v flag indicates that verification is skipped.)
8.6 Applying Software Maintenance to an HACMP Cluster
You can install software maintenance, called Program Temporary Fixes
(PTFs), to your HACMP cluster while running HACMP for AIX cluster services
on cluster nodes; however, you must stop cluster services on the node on
which you are applying a PTF. As with everything else in a cluster, applying
software fixes should be done in a controlled fashion.
(PTFs), to your HACMP cluster while running HACMP for AIX cluster services
on cluster nodes; however, you must stop cluster services on the node on
which you are applying a PTF. As with everything else in a cluster, applying
software fixes should be done in a controlled fashion.
With the method described below, you might even be able to keep your
mission-critical application up and running during the update process,
provided that the takeover node is designed to carry its own load and the
takeover load as well.
mission-critical application up and running during the update process,
provided that the takeover node is designed to carry its own load and the
takeover load as well.
The normal method of applying AIX fixes is to do the following:
1. Use the
smit clstop
fastpath to stop cluster services on the node on which
the PTF is to be applied. If you would like the resources provided by this
node to remain available to users, stop cluster with takeover so that the
takeover node will continue to provide these resources to users.
node to remain available to users, stop cluster with takeover so that the
takeover node will continue to provide these resources to users.
2. Apply the software maintenance to this node using the procedure
described in the documentation distributed with the PTF.
3. Run the
/usr/sbin/cluster/diag/clverify
utility to ensure that no errors
exist after installing the PTF. Test the fix as thoroughly as possible.
4. Reboot the node to reload any HACMP for AIX kernel extensions that may
have changed as a result of the PTF being applied.
If an update to the cluster.base.client.lib file set has been applied and you
are using Cluster Lock Manager or Clinfo API functions, you may need to
relink your applications.
If an update to the cluster.base.client.lib file set has been applied and you
are using Cluster Lock Manager or Clinfo API functions, you may need to
relink your applications.
cldare -v -M <resgroup name>:stop