Crash on Linux remote monitoring server during hub failover
to Hot Standby
The Tivoli Enterprise Monitoring Server can use a large number of file descriptors,
especially in a large environment. On UNIX and Linux systems, the maximum
number of file descriptors available to a process is controlled by user limit
parameters. To display the user limits, run the following command:
ulimit -a
The "nofiles" parameter is the number of file descriptors available to a process. For
the monitoring server process (kdsmain), the "nofiles" parameter should be set
larger than the maximum number of agents that will be connecting to the
monitoring server. If the monitoring server is unable to get file descriptors when
needed, unexpected behavior can occur, including program failures. Consider
increasing the value to 1000 file descriptors or more.
There are other user limit parameters that control how much data, stack and
memory are available to a process. For large environments, consider increasing
these memory-related user limit parameters for the monitoring server (kdsmain)
process.
Configuring the user limit parameters usually requires root access, and involves
changing system startup files which are operating system specific. Consult the
operating system manuals for information on how to configure the user limit
parameters.
HUB Tivoli Enterprise Monitoring Server quiesce prevents the
display of the data collected by the attached Tivoli Enterprise
Monitoring Agents
A HUB Tivoli Enterprise Monitoring Server has been running. A shutdown of the
Tivoli Enterprise Monitoring Server and Tivoli Enterprise Monitoring Agents on
the remote systems is in process, but the shutdown takes awhile due to abends in
the remote Tivoli Enterprise Monitoring Server. About 8 or 9 minutes go by before
the HUB Tivoli Enterprise Monitoring Server quiesces. There are a lot of remote
request communication messages in the HUB's RKLVLOG prior to the QUIESCE,
but no other signs of errors until after the abend. You cannot restart the remote
environments following the quiesce, until after the HUB environment is recycled.
The value of the MINIMUM parameter within the KDSSYSIN member of the
RKANPARU library might need to be increased if the STGDEBUG(X) or
STGDEBUG(Y) parameter is also supplied within KDSSYSIN. If the address space
controlled by this KDSSYSIN member enters a "storage quiesce" state (indicated by
a KLVxxxxx message stating that there is a storage shortage or quiesce in effect),
you should increase the value of the MINIMUM parameter and restart the address
space.
During installation of a remote Tivoli Enterprise Monitoring
Server on a Windows system, the agent support is applied,
but fails
In a multiple-level Tivoli Enterprise Monitoring Server environment, the remote
Tivoli Enterprise Monitoring Servers obtain their agent support from the hub Tivoli
Enterprise Monitoring Server. In order to apply agent support to a remote Tivoli
Enterprise Monitoring Server, the hub Tivoli Enterprise Monitoring Server must be
running and reachable by the remote Tivoli Enterprise Monitoring Server.
162 IBM Tivoli Monitoring: Troubleshooting Guide
To Next Page
Loading...
Need help?
General
