PULSE (8)
heartbeating daemon for monitoring the health of cluster nodes
SYNOPSIS
pulse [-c file |--configfile= file ] [-n|--nodaemon]
[--forceactive] [--lvs= path ] [--fos= path ]
[-v|--verbose] [--norun] [-?|--help] [--usage]
DESCRIPTION
pulse is a daemon used by lvs nodes to monitor one another's health
and to initiate the failover of the lvs or fos daemon.
It makes use of the
lvs.cf(5) configuration file for items such as length of time between
heartbeats and length of time before declaring the remote server to be dead
and initiate a failover.
pulse, which runs on the primary and backup nodes, is used by
each to determine that the other is still alive. If the backup node fails to
receive a heartbeat from the primary node after a set amount of time
(specified in lvs.cf), it will assume ownership of all virtual
service(s) being provided by the primary node. Its takeover process includes
bringing up aliased device(s) for all virtual service(s) as well as sending
out gratuitous ARPs to associate the floating IP(s) with the MAC address
of the backup node.
If and when a failed primary node returns to a functioning status, it assumes
the backup role unless specifically forced to assume
the primary role. To reduce the number of faults seen by clients and avoid the
breaking persistent client connections, you should minimize the number of
failovers.
OPTIONS
SEE ALSO
AUTHOR
Erik Troan <ewt@redhat.com>
|