Timeout when making query while purging

[aalegriadg]

I have tried to make a query during a purge of many documents from the database and I get the following timeout error:

Error running query. Reason: (timeout) The request could not be processed in a reasonable amount of time.

And the following couchdb log error:

[error] 2022-10-03T15:25:47.790110Z [email protected] <0.27022.0> aa85e646df fabric_worker_timeout map_view,'[email protected]',<<"shards/00000000-7fffffff/database.1664801850">>

It catches my attention since I am querying a partition whose documents are not included in the purge. That is, I am purging documents from partition2 and querying a view on partition1 using the get_partitioned_view_result method.

What is the cause of the error? Is it not possible to perform queries while documents are being purged? Is it due to couchdb configuration?

[nickva]

It's probably because the view shard is processing the purge requests and has to catch up. Processing a lot of purge requests is about same as the making a lot of db updates and waiting for the view to catch up.

Partition could be located on the same shard as the other partitions and views are built per-shard. So, if the shard hasn't caught up yet, it will block clients waiting the latest updates from the partitions on that shard. One way to handle it could be to split he shard range to hopefully isolate the partition to just one shard.