Hi,
I’m performing resiliency tests on my app with Crate 5.7.5 and I stumbled upon an error description which, in my opinion, does not match the Error Code returned by the HTTP Endpoint.
I’m received an error with Code 4000 with the following message :
SQLParseException[Couldn't create execution plan from logical plan because of: no such shard:
Get[sch.indicator_source | date_activated, date_created, date_deactivated, date_first_live_data, display_name, id, parent_id, meter_info, date_first_data | DocKeys{$1} | (id = $1)] (rows=1)]
It appears that the underlying error is caused by shard not being available at query time (the cluster was restarting when the query occurred).
This error should be recoverable once the cluster has properly started however and SQLParseException is not recoverable since one expects it to occur again upon a retry.
Is there a way to get a different Error code (5002 or 5003 seem appropriate) under such circumstances ?