shubhamvishu commented on PR #12183:
URL: https://github.com/apache/lucene/pull/12183#issuecomment-1716957965
@jpountz I have made some changes to the `TermStates#build` to unblock this
PR and avoid the deadlock issue happening due to executor forking into itself
by checking if its a `ThreadPoolExecutor` executor and check the executor
state(i.e. no. of threads actively executing) and only submit the task to it
when there are enough free threads to do so. i.e. we are making use of only
considering `(X-1)` threads of the thread pool(which is having fixed `X`
threads) and keeping/leaving at least **one thread** behind to avoid any
deadlock issues so that there would always be one thread available to run those
tasks("problemtic" tasks which are forking again into the executor). This
passes all the tests and also the failing seeds. Instead of keeping this PR
blocked till we address #12438 maybe we could consider this fix(short term?) to
unblock this PR but in long term we could have a recommended concurrent
executor as you mentioned in the issue #12438 to solve these issues. Let me
know what
do you think?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
