After updating ES and Kibana to 7.3.0 I have been trying to reindex old indices (v6..3.0) to new format using the Upgrade Assistant.
I succeded with a few and the I started at whole bunch (20) reindex processes.
Unfortunately I had to recycle ES which put the jobs in a Cancelled state.
But restarting the jobs does absolutely nothing.
It just reports "Reindexing" - but I can see that nothing changes when looking at the details.
My indices are rather small (< 200mb) so reindexing shoul be rather quick.
Starting reindexing processing on untouched indices doesnt work either (no progress).
I cannot identify any jobs/tasks using GET _tasks.
Can I in any way reset/cncel these jobs - or find the roort cause for this strange bhaviour?
Hm. That sounds like it's possibly a legitimate issue worth opening in GitHub.
I'm not super familiar with this part of Kibana (but I've contacted someone who is, so they may chime in). I think that it is just a thin wrapper on top of Elasticsearch's reindex API. If so, you should be able to work around this by going directly to Elasticsearch and cancelling the reindex operations.
You should be able to re-try the failed re-index in the UI. Are you able to do that? Do you get an error? Are there any errors in the Kibana server logs?
I did that reindexing to three indice one by one when I moved to 7.x
one is still working, two got so messed up. I had to retrieve data and recreate new index.
basically I read what ever field what needed from reindex put it on test machine.
delete reindex and read from test machine back to production and recreated.
No matter if I write myIndex or an indename, I get the same response:
{
"error": "no handler found for uri [/api/upgrade_assistant/reindex/MyIndex?pretty] and method [GET]"
}
{
"error": "no handler found for uri [/api/upgrade_assistant/reindex/logstash-2018.06.25?pretty] and method [GET]"
}
I can retry the failed index using the UI - but it seems that it doesn't do anything. It just claims that it is reindexing but doesn't show any progress.
I managed to solve the problems.
I downloaded Curator 5..7.6 and used that to reindex every day in each month to a monthly index.
Then I deleted all the original indicies and made a new Index Pattern in Kibana.
Now all my indecies at v7 - and furthermore i managed to reduce number of shrads from more than 4.000 to 76 (way under the new limit of 1.000).
Now I'm only missing upgrading logstash, which teased me a bit as I didn't recieve any new log-entries.
But that is a separate store, which I'll return to, if I need further assistance.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.