[
https://issues.apache.org/jira/browse/BEAM-7520?focusedWorklogId=301252&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-301252
]
ASF GitHub Bot logged work on BEAM-7520:
----------------------------------------
Author: ASF GitHub Bot
Created on: 26/Aug/19 13:24
Start Date: 26/Aug/19 13:24
Worklog Time Spent: 10m
Work Description: je-ik commented on issue #9190: [BEAM-7520] Fix timer
firing order in DirectRunner
URL: https://github.com/apache/beam/pull/9190#issuecomment-524858782
@kennknowles I removed the Optional stuff from this PR, so that it can be
focused. I have refactored the code in `StatefulParDoEvaluator` a bit, it now
enables firing multiple timers in one bundle (as long as firing any timer does
not set or delete any timer that is in the current bundle). If a timer sets
timer for time belonging in the current bundle, then all other (not yet fired)
timers are pushed back to later bundle. Reason for this is that the logic of
handling TimerUpdates lies on `WatermarkManager` and should not be performed by
`StatefulParDoEvaluator`. Thanks for the discussion, I didn't dig deeper into
the code to see, that I can get quite cleanly instance of
`DirectTimerInternals` to get the information that was needed. I think, that
now the impact on performance will be truly negligible, because overwhelming
majority of cases will fire all timers in bundle that is being processed and
only when needed, timers will be pushed back.
----------------------------------------------------------------
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:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 301252)
Time Spent: 8h 40m (was: 8.5h)
> DirectRunner timers are not strictly time ordered
> -------------------------------------------------
>
> Key: BEAM-7520
> URL: https://issues.apache.org/jira/browse/BEAM-7520
> Project: Beam
> Issue Type: Bug
> Components: runner-direct
> Affects Versions: 2.13.0
> Reporter: Jan Lukavský
> Assignee: Jan Lukavský
> Priority: Major
> Time Spent: 8h 40m
> Remaining Estimate: 0h
>
> Let's suppose we have the following situation:
> - statful ParDo with two timers - timerA and timerB
> - timerA is set for window.maxTimestamp() + 1
> - timerB is set anywhere between <windowStart, windowEnd), let's denote that
> timerB.timestamp
> - input watermark moves to BoundedWindow.TIMESTAMP_MAX_VALUE
> Then the order of timers is as follows (correct):
> - timerB
> - timerA
> But, if timerB sets another timer (say for timerB.timestamp + 1), then the
> order of timers will be:
> - timerB (timerB.timestamp)
> - timerA (BoundedWindow.TIMESTAMP_MAX_VALUE)
> - timerB (timerB.timestamp + 1)
> Which is not ordered by timestamp. The reason for this is that when the input
> watermark update is evaluated, the WatermarkManager,extractFiredTimers() will
> produce both timerA and timerB. That would be correct, but when timerB sets
> another timer, that breaks this.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
