Moreover, the impact can be critical on a running cluster (losing 2 nodes on a 3 nodes cluster can have some consequences on traffic and response time if it is on a production cluster due to the fact that the cluster will run on one node only for a moment, from the time to restart others nodes then resyncing with SST ...)
IMO, this should normally not arrived in this case or in any case in a cluster environment as it is against the logic of having a cluster if finally all nodes but one go down !?
Hi,
I strongly think that this bug should be fixed and take into consideration with a higher priority because its occurrence is increasing and it is now opened/known for a long time, a simple search on Codership Google group leads to some cases reported recently : https:/ /groups. google. com/forum/ ?fromgroups= #!searchin/ codership- team/$2Bbug$ 2F587170$ 20
Moreover, the impact can be critical on a running cluster (losing 2 nodes on a 3 nodes cluster can have some consequences on traffic and response time if it is on a production cluster due to the fact that the cluster will run on one node only for a moment, from the time to restart others nodes then resyncing with SST ...)
IMO, this should normally not arrived in this case or in any case in a cluster environment as it is against the logic of having a cluster if finally all nodes but one go down !?
Regards,
Laurent