Hi,
May i request your help, i'm facing error on ILM, request for fix please.
Thank you!
I have a similar issue. It looks like these are common problems with the setup
of ILM, which is complicated. It's easy to miss the bootstrapping step the first time.
From the following error, it appears the ILM policy created for my index was
never properly bootstrapped on the index:
[instance-0000000016] policy [app-30day-rollover] for index [app-2020.04] failed on step [{"phase":"hot","action":"rollover","name":"check-rollover-ready"}]. Moving to ERROR step java.lang.IllegalArgumentException: source alias [medengine] does not point to a write index at [...]
And yet, this time-based index has definitely rolled over, as there is now an 'app-2020.05' index which is 'hot' as well.
So my question is: How does one fix this? Both indices are giving errors and backtraces now, similar to above, every time the rollover is triggered, about every 10 min.
Some more specifics:
I created an index template, called 'app-template', with the following settings:
{
"index": {
"lifecycle": {
"name": "app-30day-rollover",
"rollover_alias": "medengine"
},
"number_of_replicas": "2",
"default_pipeline": "medengine-pipeline"
}
}
And here is the policy phase definition:
{
"policy": "app-30day-rollover",
"phase_definition": {
"min_age": "0ms",
"actions": {
"rollover": {
"max_size": "60gb",
"max_age": "30d"
},
"set_priority": {
"priority": 1
}
}
},
"version": 4,
"modified_date_in_millis": 1589683108543
}
And here is the first section of the index definition for app-2020.05
{
"settings": {
"index": {
"lifecycle": {
"name": "app-30day-rollover",
"rollover_alias": "medengine"
},
"number_of_shards": "1",
"provided_name": "app-2020.05",
"default_pipeline": "medengine-pipeline",
"creation_date": "1588291200850",
"priority": "100",
"number_of_replicas": "2",
"uuid": "XXXXXXXXXXXXXXXXXXXXXXX",
"version": {
"created": "7060299"
}
}
},
[...]
You can see the alias parameter appears to be correct, and matches the index template.
There is also an index pattern that includes the indices above: app*
It's hard to understand what to do next in this situation, without
deleting the indices with all their data and starting again. It seems
like there should be a better solution than that.
FWIW, the backtrace is pretty difficult to navigate as a jumping off point.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.
© 2020. All Rights Reserved - Elasticsearch
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.