Details
Description
Although this behaviour is as per design, this ticket is to investigate if things can be improved in fail-over scenario. There may be some complications in forcefully taking over VB ownership without a rebalance. Especially if recovery is performed soon after marking a node failed-over (without rebalance).
Attachments
Issue Links
For Gerrit Dashboard: MB-40300 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
132749,33 | MB-40300 : automatically redistribute vbs to active nodes on node failover | unstable | eventing | Status: MERGED | +2 | +1 |
133795,4 | MB-40300 : Handle race in back-to-back failover of multiple eventing nodes | unstable | eventing | Status: MERGED | +2 | +1 |
134340,3 | MB-40300 : Rename ERR_REBALANCE_ONGOING to ERR_REBALANCE_OR_FAILOVER_ONGOING | unstable | eventing | Status: MERGED | +2 | +1 |