fix panic in PeriodicJobEnqueuer #73
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
bgentry
force-pushed
the
bg-fix-periodic-enqueuer-panic
branch
from
3fc4029 to
3585375
Compare
The previous version of this code could sometimes provide a non-positive interval for Ticker.Reset, which causes it to panic. This could happen if a job was already overdue to run when `timeUntilNextRun()` was called, or as time progressed throughout its execution. There are two fixes here: 1. In `timeUntilNextRun()`, only evaluate the current time once and use that throughout so we can be sure we're not skipping safety checks. 2. Switch the `PeriodicJobEnqueuer` loop to use a `time.Timer` instead of a `time.Ticker` since we intend to reset it on every iteration anyway. Step (2) was tricky as it always is when reusing `time.Timer` in Go [1] thanks to its poor design, but I believe I've done that in a safe way here. The reusable timer is now created with 0 duration, ensuring that it will fire quickly, and then the next line always blocks until it fires. This introduces a minimal delay, but it's done to ensure that the first part of the loop (resetting the timer) can always succeed without needing to first try to stop the timer. The only other way the loop can repeat is if we've just received from the timer, so we are guaranteed to be able to safely reset it here. [1]: https://blogtitle.github.io/go-advanced-concurrency-patterns-part-2-timers/
bgentry
force-pushed
the
bg-fix-periodic-enqueuer-panic
branch
from
3585375 to
593ae13
Compare
bgentry
commented
Nov 26, 2023
Comment on lines
+171
to
+176
| // Clean up timer resources. We know it has _not_ received from the | ||
| // timer since its last reset because that would have led us to the case | ||
| // above instead of here. | ||
| if !timerUntilNextRun.Stop() { | ||
| <-timerUntilNextRun.C | ||
| } |
There was a problem hiding this comment.
I added this here to ensure we're stopping and draining the timer before returning from this loop.
brandur
approved these changes
Nov 26, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The previous version of this code could sometimes provide a non-positive interval for Ticker.Reset, which causes it to panic. This could happen if a job was already overdue to run when
timeUntilNextRun()was called, or as time progressed throughout its execution.There are two fixes here:
timeUntilNextRun(), only evaluate the current time once and use that throughout so we can be sure we're not skipping safety checks.PeriodicJobEnqueuerloop to use atime.Timerinstead of atime.Tickersince we intend to reset it on every iteration anyway.Step (2) was tricky as it always is when reusing
time.Timerin Go [1] thanks to its poor design, but I believe I've done that in a safe way here. The reusable timer is now created with 0 duration, ensuring that it will fire quickly, and then the next line always blocks until it fires. This introduces a minimal delay, but it's done to ensure that the first part of the loop (resetting the timer) can always succeed without needing to first try to stop the timer. The only other way the loop can repeat is if we've just received from the timer, so we are guaranteed to be able to safely reset it here.