summaryrefslogtreecommitdiffstats
path: root/extraconfig
diff options
context:
space:
mode:
authorJiri Stransky <jistr@redhat.com>2015-12-17 14:40:15 +0100
committerJiri Stransky <jistr@redhat.com>2016-01-04 18:15:15 +0100
commitac7467bacddb98fbb9fc9716ba56cad0d6f282de (patch)
tree121845068503cb6e2d06f99fa8cd1846c159e43a /extraconfig
parentbce5f65f2e036f6ef5232fdfc8025d7f864faa56 (diff)
Wait for cluster to settle in yum_update.sh
Occasionally we hit "Error: unable to push cib" during update. This is probably due to the fact that when we try to replace cib in yum_update.sh, services on the previous updated controller are still coming up and changing cib, and racing/conflicting with the cib push from yum_update.sh. This commit adds waiting for the cluster to settle before exiting from yum_update.sh, to avoid this kind of conflict. Also a check for cib-push success is added, to make the update fail properly instead of hanging indefinitely as we've observed with this issue. Change-Id: I953087e0e565474ac553fd57bea2459d2e3a6081 Closes-Bug: #1527644
Diffstat (limited to 'extraconfig')
-rwxr-xr-xextraconfig/tasks/yum_update.sh12
1 files changed, 11 insertions, 1 deletions
diff --git a/extraconfig/tasks/yum_update.sh b/extraconfig/tasks/yum_update.sh
index e32369e1..2d6b8cc2 100755
--- a/extraconfig/tasks/yum_update.sh
+++ b/extraconfig/tasks/yum_update.sh
@@ -24,6 +24,7 @@ update_identifier=${update_identifier//[^a-zA-Z0-9-_]/}
# seconds to wait for this node to rejoin the cluster after update
cluster_start_timeout=600
galera_sync_timeout=360
+cluster_settle_timeout=1800
timestamp_file="$timestamp_dir/$update_identifier"
if [[ -a "$timestamp_file" ]]; then
@@ -128,7 +129,10 @@ openstack-nova-scheduler"
pcs -f $pacemaker_dumpfile resource update mongod op stop timeout=100s
echo "Applying new Pacemaker config"
- pcs cluster cib-push $pacemaker_dumpfile
+ if ! pcs cluster cib-push $pacemaker_dumpfile; then
+ echo "ERROR failed to apply new pacemaker config"
+ exit 1
+ fi
echo "Pacemaker running, stopping cluster node and doing full package update"
node_count=$(pcs status xml | grep -o "<nodes_configured.*/>" | grep -o 'number="[0-9]*"' | grep -o "[0-9]*")
@@ -188,6 +192,12 @@ if [[ "$pacemaker_status" == "active" ]] ; then
fi
done
+ echo "Waiting for pacemaker cluster to settle"
+ if ! timeout -k 10 $cluster_settle_timeout crm_resource --wait; then
+ echo "ERROR timed out while waiting for the cluster to settle"
+ exit 1
+ fi
+
pcs status
else