Retain originalIndex info when rewriting FieldCapabilities requests #27761
Conversation
albertzaharovits
added
:Internal
:Search/Search
| } | ||
|
|
||
| @Override | ||
| public void writeTo(StreamOutput out) throws IOException { | ||
| super.writeTo(out); | ||
| out.writeStringArray(fields); | ||
| OriginalIndices.writeOriginalIndices(originalIndices, out); |
There was a problem hiding this comment.
this breaks backwards compatibility, you will need if based on version in both readFrom and writeTo
|
Thank you @javanna for the instant review! I have added a test of original indices for I am not sure about Can you please take another look? |
…27761) A FieldCapabilities request can cover multiple indices (or aliases pointing to multiple indices). When rewriting the request for each index, store the original requested indices.
|
@albertzaharovits @javanna @jimczi we are using sql contains 1500+ indices, and found client nodes have must serialization cost for executing sql schema check and shard search. I think it is not expected, why we need keep the original indices in above shard level request (starting from #7319), instead of just like the shard request which don't contains original indices?
|
A FieldCapabilities request can cover multiple indices (or aliases pointing to multiple indices).
The original request is rewritten for each concrete index but these new requests miss the original requested indices.
This PR mimics GetFieldMapping behavior, see GetFieldMappingsIndexRequest.