Moved 3 second delay in ovs-agent start script.

Currently, pacemaker checks status of ovs-agent by verifying its
PID after the start of an ovs-agent process, this may lead to
invalid status if the process fails to start.

This commit moves 3 second delay from the end of the 'while' cycle
of the status check to the beging, to let neutron-ovs-agent resource
set correct status

Change-Id: I1678ef82a76c6e820df3f736bc538439966e5f94
Closes-Bug: #1585824
This commit is contained in:
kkalynovskyi 2016-05-30 14:49:44 +03:00 committed by Maksim Malchuk
parent 14abded111
commit c1723ee4b8
1 changed files with 1 additions and 1 deletions

View File

@ -303,6 +303,7 @@ neutron_ovs_agent_start() {
# Spin waiting for the server to come up.
# Let the CRM/LRM time us out if required
while true; do
sleep 3
neutron_ovs_agent_monitor
rc=$?
[ $rc -eq $OCF_SUCCESS ] && break
@ -310,7 +311,6 @@ neutron_ovs_agent_start() {
ocf_log err "OpenStack (${OCF_RESKEY_binary}) start failed"
exit $OCF_ERR_GENERIC
fi
sleep 3
done
ocf_log info "OpenStack OVS agent (${OCF_RESKEY_binary}) started"