Upgrade assistant re-index helper writes off cluster

When I use the re-index helper on my 5.6.6 cluster for either the kibana or watcher index elastic will not start on nodes any more.

in the logs I get

Error injecting constructor, java.lang.IllegalStateException: index and alias names need to be unique, but the following duplicates were found [.watches (alias of [.watches-6/w4Y-TvLGSveLC8SygHm3xg])]
  at org.elasticsearch.gateway.GatewayMetaState.<init>(Unknown Source)
  while locating org.elasticsearch.gateway.GatewayMetaState

I've not run an update this is just running the index helper and then restarting the nodes. If I then go on each node and under the data directories (I run 2) delete the folders /data/elasticsearch/nodes/0/indices/TvLGSveLC8SygHm3xg then the node will restart but repeate the error on further restarts. If I delete it on all nodes and do a full restart then it will all load back up but the kibana index gets re-created as a v5.x index and I'm told I need to upgrade it again by the upgrade assistant.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.