Details
-
Bug
-
Resolution: Fixed
-
Major
-
2.1.0, 2.2.0
-
Security Level: Public
-
Untriaged
Description
steps:
1. set autofailove 30 secs
2. set email alerts settings
3. 2 nodes in cluster (112, 113)
4. stop node 113
observation :
get error in logs as expected
Could not auto-failover node ('ns_1@10.3.121.113'). Cluster was too small, you need at least 2 other nodes.
auto_failover004 ns_1@10.3.121.112 13:47:37 - Mon Apr 15, 2013
Node 'ns_1@10.3.121.112' saw that node 'ns_1@10.3.121.113' went down. Details: [
5. restart 113 and rebalance out from cluster
Here I played with three nodes in clusters, but do not think that these steps are important
6. 1 node 112 in cluster, set again autofailove 30 secs and email alerts settings( test mail)
7. rebalance in 113 node and the stop it
result:
get only message that node 113 went down
Node 'ns_1@10.3.121.112' saw that node 'ns_1@10.3.121.113' went down. Details: [{nodedown_reason, connection_closed}
]
I believe that the user should again get the message "Could not auto-failover node ('IP'). Cluster was too small, you need at least 2 other nodes."
Attachments
For Gerrit Dashboard: MB-8089 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
32622,3 | MB-8089 Resurrect per node uuid. | master | ns_server | Status: MERGED | +2 | +1 |
32623,3 | MB-8089 Pass node name coupled with UUID to auto_failover_logic. | master | ns_server | Status: MERGED | +2 | +1 |
32624,3 | MB-8089 Look at nodes set instead of size in auto_failover_logic. | master | ns_server | Status: MERGED | +2 | +1 |