Fix indefinite client hang after connection reset #29
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
We discovered a problem in this library that causes the client to hang indefinitely. This patch makes sure that the locking part (`net.textproto.Pipeline.StopRequest`) is executed whenever the function returns. Steps to reproduce: 1. Establish a normal connection with `beanstalkd` and keep the connection open 2. Reserve a job through `TubeSet.Reserve(...)` 3. Exit the server 4. Reserve a job through `TubeSet.Reserve(...)`. The expected error is: `reserve-with-timeout: read tcp srcip:srcport->dstip:dstport: read: connection reset by peer` 5. Reserve a job through `TubeSet.Reserve(...)`. The expected error is: `reserve-with-timeout: write tcp srcip:srcport->dstip:dstport: write: broken pipe` After step 3, `t.Conn.cmd(...)` (in `tubeset.go:30`) is executed without error, because writes to a half-closed socket are accepted. Step 4, then, fails with the error mentioned. In step 5, the write fails and `t.Conn.cmd(...)` returns early in `conn.go:76`. However, `c.c.StartRequest(r.id)` is not properly ended and this causes jobs with an `id` larger than `r.id` to wait indefinitely to be processed. Cc: @Minnozz Fixes: beanstalkd#21
|
To summarize: The client will sometimes block indefinitely because of a bug where |
xuri
referenced
this pull request
in xuri/aurora
May 28, 2018
https://github.com/kr/beanstalk/pull/29; - go test has been updated
mrngm
changed the title
c.c.EndRequest(r.id) when c.cmd(...) returns
|
Thank you very much! |
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Outdated suggestions cannot be applied.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.
We discovered a problem in this library that causes the client to hang
indefinitely. This patch makes sure that the locking part
(
net.textproto.Pipeline.StopRequest) is executed whenever the functionreturns.
Steps to reproduce:
beanstalkdand keep the connection openTubeSet.Reserve(...)TubeSet.Reserve(...). The expected error is:reserve-with-timeout: read tcp srcip:srcport->dstip:dstport: read: connection reset by peerTubeSet.Reserve(...). The expected error is:reserve-with-timeout: write tcp srcip:srcport->dstip:dstport: write: broken pipeAfter step 3,
t.Conn.cmd(...)(intubeset.go:30) is executed without error,because writes to a half-closed socket are accepted. Step 4, then, fails with
the error mentioned. In step 5, the write fails and
t.Conn.cmd(...)returnsearly in
conn.go:76.However,
c.c.StartRequest(r.id)is not properly ended and this causes jobswith an
idlarger thanr.idto wait indefinitely to be processed.Cc: @Minnozz
Fixes: #21