diff options
| -rwxr-xr-x | scripts/rabbitmq-server-ha.ocf | 39 |
1 files changed, 37 insertions, 2 deletions
diff --git a/scripts/rabbitmq-server-ha.ocf b/scripts/rabbitmq-server-ha.ocf index 9b3acd9803..c74525c936 100755 --- a/scripts/rabbitmq-server-ha.ocf +++ b/scripts/rabbitmq-server-ha.ocf @@ -812,10 +812,45 @@ get_master_name_but() done } +erl_eval() { + local fmt="${1:?}" + shift + + ${OCF_RESKEY_ctl} eval "$(printf "$fmt" "$@")" +} + # Returns 0 if we are clustered with provideded node is_clustered_with() { - get_running_nodes | grep -q $(rabbit_node_name $1); + local LH="${LH}: is_clustered_with: " + local node_name + local rc + node_name=$(rabbit_node_name $1) + + local seen_as_running + seen_as_running=$(erl_eval "lists:member('%s', rabbit_mnesia:cluster_nodes(running))." "$node_name") + rc=$? + if [ "$rc" -ne 0 ]; then + ocf_log err "${LH} Failed to check whether '$node_name' is considered running by us" + # XXX Or should we give remote node benefit of a doubt? + return 1 + elif [ "$seen_as_running" != true ]; then + ocf_log info "${LH} Node $node_name is not running, considering it not clustered with us" + return 1 + fi + + local seen_as_partitioned + seen_as_partitioned=$(erl_eval "lists:member('%s', rabbit_node_monitor:partitions())." "$node_name") + rc=$? + if [ "$rc" -ne 0 ]; then + ocf_log err "${LH} Failed to check whether '$node_name' is partitioned with us" + # XXX Or should we give remote node benefit of a doubt? + return 1 + elif [ "$seen_as_partitioned" != false ]; then + ocf_log info "${LH} Node $node_name is partitioned from us" + return 1 + fi + return $? } @@ -1607,7 +1642,7 @@ node_health_check_local() { return $OCF_ERR_GENERIC elif [ "$rc_timeouts" -eq 1 ]; then ocf_log info "${LH} node_health_check timed out, going to retry" - return $OCF_ERR_GENERIC + return $OCF_SUCCESS fi if [ "$rc" -ne 0 ]; then |