Juan Pedro Paredes Caballero <[email protected]>

HA sync clones periodicaly 2 nodes in a pacemaker+corosync cluster
It should run periodically to sync cluster
It takes care to not to load the cluster, limiting bandwidth, io and cpu
This script must be modified to fit your needs.
You should use it to to sync debian system software, not service data.
You can use it with other flavours of UNIX but you need to adjust 
the exclude paths to not sync the single node information.
This script must be modified to fit your needs.
For service data like a database use DRBD o something like that.
Please look at RESOURCE,MYEXCLUDE,ECLUDE,SERVCCES and OPTS variables

hasync expect to find fairsync in /usr/local/bin/

hasync uses fairsync as a wra
#This the HA resource to stablish who is primary and who is standby
RESOURCE="cluster-ip"
#Exclude
MYEXCLUDE="/etc/mysql/conf.d  /var/lib/mysql"
#This variable permit to stop services while sync is in progress
SERVICES=""
#We use the fairsync wrapper to rsync to minimize the use of IO and CPU
OPTS='--sockopts=SO_SNDBUF=2000000,SO_RCVBUF=2000000 --rsync-path="/usr/local/bin/fairsync"'