ServiceGuard Manager Version A.02.00 Release Notes, June 2002
32 Chapter1
ServiceGuard Manager Version A.02.00 Release Notes
Patches and Fixes in this Version
DOWN clusterwide. If a package is clearly down on any node, it must
be down on all nodes. However, it is possible for a package to be
starting or halting for a short time on an individual node, and this
state is not shown in ServiceGuard Manager.
• What is the workaround? There is no work-around.
JAGad86950
• What is the problem? On a few rare occasions in testing, the map was
not drawn when the selection in the tree changed.
• What is the workaround? Change the focus of the tree and then
change it back to redraw the map.
JAGad87580
• What is the problem? Late error dialog when connection is dropped.
When the connection is dropped, the connection status properly
indicates the condition but the dialog box that reports the problem is
not displayed until the second attempt to refresh the data fails.
• What is the workaround? Click on the Refresh button if the status of
the connection is ever in doubt.
JAGad88000, 90300
• What is the problem? Packages with no failover nodes show a
warning, even if they were configured with no failover node.
ServiceGuard Manager displays warnings when a package has no
failover. However, some nodes are configured to have only one node,
and to not fail over. For example, OPS packages are always configured
without failover nodes.
In the map, the package will have a yellow border and have a
no-failover badge beside it. In addition, this will cause the cluster to
also have the yellow border and badge.
This continual warning state may cause you to ignore problems with
other packages in the cluster. If there is a node problem, the cluster
usually gets a yellow border. In this cage, however, the cluster is
already yellow so you won’t see a change in the cluster icon.
• What is the workaround? If a cluster has single-node packages in it,
keep that cluster’s map always open, so you can see other packages.