nickva commented on a change in pull request #3969:
URL: https://github.com/apache/couchdb/pull/3969#discussion_r836807224
##########
File path: src/fabric/src/fabric_view.erl
##########
@@ -469,8 +481,14 @@ fix_skip_and_limit(#mrargs{} = Args) ->
{CoordArgs, WorkerArgs} =
case couch_mrview_util:get_extra(Args, partition) of
undefined ->
- #mrargs{skip = Skip, limit = Limit} = Args,
- {Args, Args#mrargs{skip = 0, limit = Skip + Limit}};
+ #mrargs{skip = Skip, limit = Limit, max_view = Max} = Args,
Review comment:
The `max_view` field here I think is always 268435456? In that case
adding or not adding an offset to it probably doesn't make a large difference,
as the value is already huge. But this also points to a problematic spot. When
previously we limited our per-node emits when user passed in a limit, we won't
be doing that any longer and start emitting all the rows. I _think_ it should
be ok given that we have the rexi streaming replies (acks) and cleanup happens
properly. But if they don't happen properly then this could add significant
processing load on the workers.
--
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]
