This daemon is part of the BProc package and is installed by default. It is the controller and message/IO manager for all the compute nodes and must be running for the cluster to function.
It is started in the /etc/rc.d/init.d/beowulf script, along with other BProc daemons, and forks a copy of itself for IO forwarding. With normal cluster operation there should be 2 PIDs for bpmaster. Type ps -x |grep bpmaster to check.
The daemons may be restarted at any time with /etc/rc.d/init.d/beowulf restart, but note that this will cause all nodes to reboot. During normal operations, you should use the beosetup GUI or /etc/rc.d/init.d/beowulf reload to enable cluster configuration changes. The beosetup GUI will display "master down" for all nodes when the daemon is not running. Bpmaster uses /var/log/messages to report cluster events such as configuration changes and errors.
The following options are available to the bpmaster program.
Print the command usage message and exit. If -h is the first option, all other options will be ignored. If -h is not the first option, the other options will be parsed up to the -h option, but no action will be taken.
Print the command version number and exit. If -v is the first option, all other options will be ignored. If -v is not the first option, the other options will be parsed up to the -v option, but no action will be taken.
Start the program in debug (verbose) mode. bpmaster will not daemonize, and all information and error messages will go to stdout. This information is useful when the daemon exits abnormally during operation as the information is not mixed in with the normal /var/log/messages
Increase verbosity level. This may be specified multiple times.
Ignore interface version mismatch. This can be dangerous.
Specifies a different configuration file is to be used. The default is set to /etc/beowulf/config. This option is for debug and development, and does not work in conjuction with beosetup and other GUI tools. This option is not recommended for normal use.
Log master and node BProc messages to the indicated file. This information is intended for BProc debugging, and should not be enabled unless requested by a SCYLD support engineer. This file grows in size rapidly depending of the number of nodes, aproximatly 2 megabytes/minute with six nodes.
Don't start as daemon.
[root@cluster /root]# bpmaster -d
Start the daemon using the startup script.
[root@cluster /root]# /etc/rc.d/init.d/beowulf start Configuring network interface (eth1): [ OK ] Loading modules: [ OK ] Setting up libraries: [ OK ] Starting bpmaster: Starting beoserv: Starting recvstats: Starting sendstats: