| Scyld ClusterWare HPC: Reference Guide | ||
|---|---|---|
| << Previous | Next >> | |
This node argument should list one of the compute nodes, not the master. The master is considered to be always up.
On error, it will return -1 and set errno appropriately.
The possible states are:
bproc_node_downThe node is not connected to the master daemon. It may be off or crashed or not far enough along in its boot process to connect to the master daemon.
bproc_node_unavailableThe node is running but is currently unavailable to users. Nodes are only in this state if set that way explicitly by the administrator.
bproc_node_errorThere is a problem with the node. Nodes are assigned this state if booting is unsuccessful.
bproc_node_upThe node is up and ready to accept processes. This is the only state in which non-root users can send jobs to the node.
bproc_node_rebootThe node was told to reboot and has not come back up yet.
bproc_node_haltThe node was told to halt and is still down.
bproc_node_pwroffThe node was told to power off and is still down.
bproc_node_bootThe node is in the process of coming up (running the node_up script).
| << Previous | Home | Next >> |
| bproc_nodenumber | Up | bproc_numnodes |