Serviceguard Version A.11.16 Release Notes, 2nd Edition, September 2004

Serviceguard Version A.11.16 Release Notes, Second Edition
Known Problems and Workarounds
Chapter 152
JAGae87101: LOG_PERIODIC messages should not be
logged at log level 1
• What is the problem? Because periodic messages show up very often,
enabling this sort of debug logging will fill up syslog.log very
quickly.
• What is the workaround? There is no workaround.
JAGae62205: Serviceguard package cannot be
restarted after hardware monitoring is re-enabled.
• What is the problem? When hardware monitors are disabled using
the monconfig tool, all associated hardware monitor requests are
removed from the persistence files; when hardware monitoring is
re-enabled, these requests are re-created. However, hardware
monitor requests that were created using SAM, or established when
Serviceguard is started, are related to the psmmon hardware monitor;
these requests are not re-created.
Here is one example:
1. A package is configured in SAM with a hardware monitor
request. The cluster starts, Serviceguard creates a request, and
the package starts.
2. Later, an administrator stops hardware monitoring, using
monconfig -K. All monitors stop, and all related requests are
removed. The package halts, to move to another node.
3. The administrator re-enables the hardware monitors using
monconfig -E. However, the requests that were created outside
of monconfig are not re-created. The package can no longer start,
because the monitor request is lost.
• What is the workaround? If you have created a hardware monitor
request without using the monconfig tool and then you stop the
monitors, you must re-start the requests yourself. Here are two
different ways to do that.
— While the cluster is running and after hardware monitoring has
been re-enabled, manually re-add the monitor requests using the
cmstartres command. Then start the package.