Repository navigation
TypeError: Cannot read property 'remove' of undefined at close (cluster.js:497:15) #3843
Description
I am seeing these blowups on Travis.
I will add some portions of the build log.
Setting environment variables from .travis.yml
$ export TRAVIS=travis
$ export CXX=g++-4.8
6.56s$ nvm install 5.0
######################################################################## 100.0%
Checksums empty
Now using node v5.0.0
$ node --version
v5.0.0
$ npm --version
3.3.6
$ nvm --version
0.23.3
During the testing I am seeing the following.
Beginning prism shutdown
Beginning store shutdown
Beginning store shutdown
cluster.js:497
if (handle.remove(worker)) delete handles[key];
^
TypeError: Cannot read property 'remove' of undefined
at close (cluster.js:497:15)
at Worker.onmessage (cluster.js:431:7)
at ChildProcess.<anonymous> (cluster.js:713:8)
at emitTwo (events.js:92:20)
at ChildProcess.emit (events.js:172:7)
at handleMessage (internal/child_process.js:686:10)
at Pipe.channel.onread (internal/child_process.js:440:11)
1) "before all" hook for "should still download content"
And another
cluster.js:497
if (handle.remove(worker)) delete handles[key];
^
TypeError: Cannot read property 'remove' of undefined
at close (cluster.js:497:15)
at Worker.onmessage (cluster.js:431:7)
at ChildProcess.<anonymous> (cluster.js:713:8)
at emitTwo (events.js:92:20)
at ChildProcess.emit (events.js:172:7)
at handleMessage (internal/child_process.js:686:10)
at Pipe.channel.onread (internal/child_process.js:440:11)
2) "after all" hook for "should still download content"
I am not sure what the internals are of what are causing this. However, I can note that on my Windows machines with Node 5.0 I am not getting these errors.
These crashes are happening randomly for me. If I restart my tests they happen in a different place each time.
It only happens during cluster shutdown.
I am doing a build on multiple node versions to see if I can figure out where this is spawning from.
Activity
nullivex commentedon Nov 16, 2015
Okay, I will try to add some environmental states here.
I am using the cluster module through a module called https://www.npmjs.com/package/infant which I have been using since Node 0.10 without issue.
I am unsure if there is something in the cluster handling causing this error. I googled for it and couldn't find anyone else seeing this blow up.
I dont see these errors in the unit tests for infant so I curious about what exactly is causing these strange errors.
nullivex commentedon Nov 16, 2015
Okay nevermind.
I found the fix against this error here: 33827e3
I guess it just hasnt been released yet.
nullivex commentedon Nov 16, 2015
Any idea when this fix will be released?
yosuke-furukawa commentedon Nov 16, 2015
You would be better to wait for release v5.1.0.
#3736
According to the PR, we will release the version soon.
Fishrock123 commentedon Nov 16, 2015
v5.1.0 will release this week, probably tomorrow. It has been delayed due to some jenkins issues: #3709
Closing this since the fix itself has already landed.