Fix replace conflict strategy #314
Closed
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
After initially failing to get this feature working, I found that the strategy class wasn't in the hash of strategies. The other thing this PR addresses is removing the client middleware is loaded on the server. The client middleware checks if a lock is in place and enqueues in its absense. Client middleware is only hit when the server enqueues to itself which happens when sidekiq takes scheduled items and enqueues them on the work queue. I was trying to use the `:until_executing` lock along with `:replace` strategy, but the server would just take the job and drop in `Client::Middleware#call` because the lock only gets released during the server middleware call. All ears if the client middleware on the server is necessary for whatever other reason. All the tests pass without it (bar the test which just checks the presence of it, which I fixed).
|
Hi @RoryDH, thank you for the contribution. The problem with this PR is that if a job has a nested job that requires uniqueness that nested job will be pushed regardless. I'll see about maybe pushing it directly to redis using Lua or some such. |
Merged
|
If you have a look in #315 that should be enough to make the replace work. The thing is that what the replace strategy does is remove the other job from the queue, scheduled queue or retry queue and then also remove the digest and locks. Now the client should be able to push this job again. It has nothing to do with executing the job. That is done from the server process while it is locked from the client push. |
|
Great thanks for explaining and making those changes! |
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.
After initially failing to get this feature working, I found that the strategy class wasn't in the hash of strategies.
The other thing this PR addresses is removing the client middleware is loaded on the server.
The client middleware checks if a lock is in place and enqueues in its absense. Client middleware on the server however is only hit when the server enqueues to itself which happens when sidekiq takes scheduled items and enqueues them on the work queue.
I was trying to use the
:until_executinglock along with:replacestrategy, but the server would just take the job and drop inClient::Middleware#callbecause the lock only gets released during the server middleware call.All ears if the client middleware on the server is necessary for whatever other reason.
All the tests pass without it (bar the test which just checks the presence of it, which I fixed).