Github user davisp commented on the pull request:
https://github.com/apache/couchdb-mem3/pull/17#issuecomment-189365989
Yeah, not so much. Search must just have a bug in its shard replacement
logic. If a node is in maintenance mode then the RPC worker dies and the
coordinator should attempt to replace it. Checking the dreyfus code the RPC
worker should die here:
https://github.com/cloudant/dreyfus/blob/master/src/dreyfus_rpc.erl#L115
And that should theoretically be handled and replaced here:
https://github.com/cloudant/dreyfus/blob/master/src/dreyfus_fabric.erl#L68
Granted there may be a bug in either not supplying replacements or a
failure in how we handle the replacement. I'd start by writing a test for that
bit and working up to the bug you're seeing. Caching info in mem3 is less than
ideal because caching is hard.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
