Github user rnewson commented on the pull request:
https://github.com/apache/couchdb-couch-mrview/pull/39#issuecomment-177162214
agree with davisp. For this to be the right fix we would _never_ close a
file after compaction, rather it seems like a subset.
Locally, OSX, I verified that compacting a view shard causes the fd to
change for the file in question and the old fd no longer appears in lsof output.
Next step is to patch a cloudant production cluster with additional logging
so we can find what must be a race condition. Log when we finish compaction (I
think we have this one already), log when we decide whether or not to close an
unmonitored couch_file (in either case).
---
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.
---
