[HttpCache] Hit the backend only once after waiting for the cache lock #60502
Conversation
This file contains hidden or 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
|
Test failures seem unrelated |
mpdude
commented
May 21, 2025
Comment on lines
+222
to
+226
| try { | ||
| $response = $this->lookup($request, $catch); | ||
| } catch (CacheWasLockedException) { | ||
| $response = $this->lookup($request, $catch); | ||
| } |
There was a problem hiding this comment.
This allows to restart the process only once.
In fact, the old code would wait for the lock to be released, but only try to acquire it later on. That could, in theory, allow for repeated cycles without actually reaching the 503 condition in lock().
Maybe it is even better this way to fail with the exception instead of silenty retrying over and over again?
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Outdated suggestions cannot be applied.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.
When the
HttpCachehas to wait for a lock held by another, concurrent process, thelock()method wants to make sure we continue operation based on the most recent cache entry, possibly updated by the concurrent process.So, after waiting for lock release, it calls
lookup()to obtain this cache entry. This is, in fact, a reentrant call up into the current call stack.Having
lookup()multiple times on the call stack opens up a way to call the backend for validation multiple times. I have observed this in practice at least in combination with theno-cachecache-control header, causing surprising side effects™️ ✨.Also without
no-cacheyou can get strange-looking cache traces likestale, valid, store, fresh. Those occur only when concurrent locking is a issue.I am not super happy with using an exception for a control flow issue like this. But, rolling back to
lookupseems to be the most sensible decision for me, and using special return values to indicate this condition isn't really pretty either.