Failed to Re-index before 5 -> 6 Upgrade, now what?

On one of my ES clusters, I failed the step to re-index before upgrading from 3.3 to 3.5.6 to 6.0.1 so naturally my data node fails to start citing errors about being unable to upgrade mappings for various indexes.

Is there anything I can do manually or am I just hosed on this cluster?

Solved my own problem by un-instaling version 6.0.1 from all the nodes and replacing it back with 5.6.5

This worked in my situation because the data nodes storing the bulk of the information were never able to come up online in version 6, so effectively still held their data in the previous 2/5 format. There was some re-indexing that had to be done for the Kibana node because that one did upgrade successfully.

Overall, need to spend a lot more time considering the upgrade to version 6 and the various implications and breaks. Might be time to just re-build from scratch in 6 and re-index the information over from the previous cluster.

Since you're back on Elasticsearch 5.6, you can try if using the XPack Upgrade Assistant that comes with XPack for Elasticsearch/Kibana 5.6 allows you to finish the upgrade successfully.

We recommend using this tool to upgrade the .security and .kibana indices (and any other indices flagged up by it), prior to upgrading the first node to 6.x Also inspect and act on any other issues flagged by the tool. For further information, please see the upgrade documentation.

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