For the most part the Event Management subsystem runs itself without requiring administrator intervention. However, on occasion, you may need to check the status of the subsystem, or add or change some of the configuration data.
This section contains the procedures that you need to do these tasks, which include:
You can display the operational status of the Event Manager daemon by issuing the lssrc command.
On the control workstation, enter:
lssrc -l -s haem.domain_name
where domain_name is the name of the domain of interest.
On a node, enter:
lssrc -l -s haem
In response, the lssrc command writes the status information to standard output. The information includes:
For information on these flags, see the haemtrcon command in PSSP Command and Technical Reference.
Daemon connected to group services: TRUE/FALSE Daemon has joined peer group: TRUE/FALSE Daemon communications enabled : TRUE/FALSE
The peer group state includes the EMCDB version string and the peer group security state. The peer group security state is the keyword SEC, NOSEC, or NOSECSUPPORT. The keyword might have a suffix, but it can be ignored. If the peer group is established by a pre-PSSP 3.2 version of the EM daemon, no security keyword is present.
A resource monitor may have multiple executing instances, the number of the resource monitor instance is specified in the Inst column. The connection type (C=client, S=server, I=internal) is found in the Type column. The connection status is indicated by the FD column; if the file descriptor is greater than or equal to 0, a connection is open. If a resource monitor has a shared memory segment used to transfer information to the Event Manager daemon, it has a shared memory ID greater than or equal to 0 in the SHMID column. The process ID of the resource monitor is listed in the PID column; it is interpreted as follows:
The Locked column indicates whether or not a resource monitor is locked and the current count of start attempts and successful connections, in the form mm/nn, where mm is the count of start attempts and nn is the count of successful connections.
If a resource monitor has more than one instance, information is present in the PID and Locked columns only for instance number 0. However, the count of successful connections is for all instances of the resource monitor.
This lists the status of peer daemons by node number, in node number order. Note that this list only includes peer daemons that have joined the peer group since the local daemon started.
Following the node number are two characters. If both characters are S, the specified node is the number of the node where this daemon is running. Otherwise, the characters can take on values as follows.
The first character is I or O where:
The second character is either A or R, where:
The default configuration data supplied for the PSSP is normally loaded into the SDR and compiled into its binary format automatically by the haemctrl script. However, if resource monitors supplied by other IBM products or by third parties are installed on the SP system, you must load the configuration data supplied with the resource monitors into the SDR and activate it. To do this:
If you are creating a new file, use the format specified in the man page for the haemloadlist file.
haemloadcfg new_loadlist
where new_loadlist is the path name of the load list you previously created or identified.
With an optional flag, the haemloadcfg command can replace existing objects (identified by their key attributes) in the SDR. If you want to change an object that already exists, do the following:
For PSSP configuration data, the default configuration data is in the /usr/sbin/rsct/install/config/haemloadlist file. You must copy this file to another file and make the changes in the copy. If you have non-PSSP or third-party resource monitors, the object will be loaded from another load list file.
Do this step for as many objects as you are changing.
In response, the haemloadcfg command replaces each object in the SDR matched by an object in the load list file.
When you have added or changed the Event Management data in the SDR, you must activate it by recompiling the EMCDB and stopping and restarting the Event Manager daemons. To do this:
haemcfg
The output is placed in the staging directory in a file whose name indicates the system partition.
Issue the haemctrl -k command on the control workstation and on each of the nodes in the system partition.
You can use the dsh or Sysctl commands to run the command on multiple nodes from the control workstation. For more information on using these commands, see Using dsh to run parallel management commands.
On the control workstation, issue the lssrc -s haem.domain_name command. On the nodes, issue the lssrc -s haem command.
You can use the dsh or Sysctl commands to run the command on multiple nodes from the control workstation. For more information on using these commands, see Using dsh to run parallel management commands.
The status of each daemon should indicate that it is inactive.
Issue the haemctrl -s command on the control workstation and on each of the nodes in the system partition.
You can use the dsh or Sysctl commands to run the command on multiple nodes from the control workstation. For more information on using these commands, see Using dsh to run parallel management commands.
To change the limit on the number of resource variable instances accepted by the Event Management subsystem for any resource variable class, do the following:
EM_Resource_Class
rcClass="IBM.PSSP.VSD"
rcResource_monitor="IBM.PSSP.harmld"
rcObservation_interval="60"
rcReporting_interval="10"
EM_Resource_Class
rcClass="IBM.PSSP.VSD"
rcResource_monitor="IBM.PSSP.harmld"
rcObservation_interval="60"
rcReporting_interval="10"
rcInstance_limit="5000"
to limit the number of Recoverable Virtual Shared Disk resource variable instances to 5000.