[GitHub] [couchdb] chewbranca commented on pull request #3161: WIP: Workaround dirty schedulers in run_queue stats

[GitBox <gi...@apache.org>]

chewbranca commented on pull request #3161:
URL: https://github.com/apache/couchdb/pull/3161#issuecomment-696320444


   > Why not do the version that includes the dirty CPU schedulers given that we're doing dirty IO schedulers?
   
   @davisp I'm assuming you meant the opposite, I'm using the version with dirty CPU schedulers since we're not using IO schedulers.
   
   More to the point, I'm using `statistics(run_queue_lengths)` [1], which returns the list or normal scheduler queues followed by a single entry for the dirty CPU queue. This works for our immediate needs and what the PR does. There's also `statistics(run_queue_lengths_all)` [2] which returns the list of normal scheduler queues followed by an entry for the dirty CPU queue and another entry for the IO dirty queue.
   
   There isn't any dirty IO in Erlang yet, and we're not using any either, so I didn't include that. I think it does make sense to report the dirty IO queues as Erlang will switch to using them at some point in the future, however, I did not do that in this PR as the `statistics(run_queue_lengths_all)` statistic didn't land until OTP-20, and we still have partial support for OTP-19.
   The `statistics(run_queue_lengths)` statistic landed in OTP-18.3 so we're free to use that. I'm not opposed to using `statistics(run_queue_lengths_all)`, but if we're going to do that I think we should drop support for OTP-19 in the process.
   
   [1] http://erlang.org/doc/man/erlang.html#statistics_run_queue_lengths
   [2] http://erlang.org/doc/man/erlang.html#statistics_run_queue_lengths_all


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org