sceptic
Check if compute host (nodes) are ailve or not.
More...
sceptic [ OPTIONS ] HOSTGROUP [ COMMAND ]
sceptic tries to check if compute hosts specified with the HOSTGROUP argument are up or down in a parallel way. If a compute host responds normally with the Linux ping(1) command, the SCore scping(8) command, and the Linux uptime(1) command, then the host is considerd to be up and running. Since the scping(8) command tries to connect with the SCore scoutd(8) process on the compute host, failure of the connection may not mean the failure of the compute host.
The Linux uptime(1) command can be any other Linux command by specifying the COMMAND argument and followed by its paramete(s). This command and the default uptime(1) command is spawned using the Linux rsh(1) command. See also the SCORE_RSH environment variable section below.
List alive hosts. This option can not be combined with the -d or -D option described below. Normally the sceptic commands lists down hosts without this option. The sceptic commands succeeds if every compute hosts is alive, if the -d option below is not specified.
List down hosts. The sceptic commands succeeds if every compute hosts is down. This option can not be combined with the -a option above.
List down hosts prefixed by tilde (~) symbol, so that the output can be used as the defect list of the scorehosts.db(5) file. This option can not be combined with the -a option above.
Number of parallelism.
List hosts with newline
Not to output summary message
Number of retry. Default is 10 times. If the number is the string of inf, then the sceptic command repeat checking until it succeeds.
Short hostname is output
Verbose mode. Individual host status is printed in real time.
Remote shell command instead of the Linux rsh command.
scorehosts(1), scoutd(8), scping(8).
- CREDIT
- This document is a part of the SCore cluster system software
developed at PC Cluster Consortium, Japan.
Copyright (C) 2003-2008 PC Cluster Consortium.