Hello Folks,
We applied a stemming override to one of our indexes via rules_path. We encountered a problem when the file pointed to in the rules_path was not found on all nodes in the cluster (someone did not copy it properly). This aborts the _SETTINGS call for the index leaving it closed.
Naturally fixing this would be a matter of ensuring the missing file on that one node of the cluster exists and reopening the index. But, if some reason the group that manages this for us cannot find out which node has the missing file, we thought we'd be able to execute _SETTINGS api to remove the RULES_PATH from the filter and analyzers, unfortunately ES doesn't allow this. It responds like:
{
"error": {
"root_cause": [
{
"type": "remote_transport_exception",
"reason": "[NODE_NAME][##.##.##.###:9300][indices:admin/settings/update]"
}
],
"type": "illegal_argument_exception",
"reason": "IOException while reading rules_path: PATHNAME\config\analysis\english_stemmer_override.txt",
"caused_by": {
"type": "file_not_found_exception",
"reason": "PATHNAME\config\analysis\english_stemmer_override.txt (The system cannot find the file specified)"
}
},
"status": 400
}
Is there some way around this to enable a rollback of the originally aborted command?
Anyone know why ES wants to resolve the state of this, before allowing us to back it out?
Thanks, Mark