Problematic socket.parser.resume() error #3508
Comments
silverwind
added
http
|
/cc @indutny |
|
I think adding single guard should be enough to cover this. Writing a regression test for it may be a bit trickier, though. Going to take a look tomorrow. |
|
Hi. We are launching the website very soon. Have you found the source of the error? Could you tell me how to fix this issue? When can I expect to get the solution? |
|
The fix is very easy, but the test unfortunately takes a bit more time. I'll continue working on it tomorrow. Sorry for delay! |
|
@binarybehemoth are you sure that this stack trace is from 4.2.0? Here is 128 line from https://github.com/nodejs/node/blob/v4.2.0/lib/_http_outgoing.js#L128 Doesn't seem to match. I guess the version is 4.1.0? |
|
So sorry. You are right. That stack trace was obtained before I upgraded node from v4.1.2 to v4.2.1, but the problem persisted after the upgrade. Here is the stack trace from v4.2.1: TypeError: Cannot read property 'resume' of null |
|
@binarybehemoth we have a fix, but it will take some time to test and make sure that there are no other regressions. Please stay tuned, sorry for delay! |
|
Great. It is good to know you have fixed the issue. Please release the new version as soon as possible. The current bug is really annoying. |
|
Has the issue been addressed? I tried using v5.0.0 and the problem was still there. When can I obtain the fixed version? |
|
@binarybehemoth the fix has not yet landed anywhere. However it should land soon, I'll keep you posted. Sorry for delay! |
|
Is there any update on this? |
|
@kbernst30 we're on it. Soon, think within two weeks. |
|
Hey, I know this has already been reported, and that people are working on it, but it's been two weeks without an update. This bug has been causing unbelievable havoc for me. http is the beating heart of Node in web deployment. I need to be able to rely on it. It's been crashing my server every few hours, and sometimes up to multiple times a second. |
|
i'm having this issue too after i upgraded node from 0.12.7 to 4.2.2, any update on this? |
|
any update on this? my server has been crashing multiple times a day... |
|
Just discovered this issue as well. Is there a date when we can expect this to land? Is there something that we can do to help push this out? |
|
I don't know about you but I got used to it |
|
A fix will be released later this week. |
|
This might be a stupid question but how do I inject this fix into my existing project (that uses v4.2.2), should I reinstall node v4.2.2? |
|
@artofspeed you should grab 4.2.3 http://nodejs.org/en/blog/release/v4.2.3/ |
|
Please be aware that this bug is related to CVE-2015-8027 and it's important that you upgrade to 4.2.3 or 5.1.1. |
Always check that socket still has the parser. It may be destroyed interim, and we may end up with an uncaught exception. Fix: nodejs#3508 PR-URL: nodejs-private/node-private#5 Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl> Reviewed-By: James M Snell <jasnell@gmail.com>
Always check that socket still has the parser. It may be destroyed interim, and we may end up with an uncaught exception. Fix: nodejs/node#3508 PR-URL: https://github.com/nodejs/node-private/pull/5 Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl> Reviewed-By: James M Snell <jasnell@gmail.com>
This bug has been crashing my node.js server for the past one week. On average it seems to occur every 12 hours. Normally when an error occurs in node.js, node.js will tell you the line number and your source file that lead to the error. However, this error is so weird that it does not let me know where the error originates from in my source codes. I can't even find _http_server.js within the entire project and node modules. As such, I have been unable to pinpoint and eliminate the bug.
Without the stack trace module 'longjohn', I would get the following error message:
socket.parser.resume()
cannot read property resume of null (_http_server.js:442)
......
After I installed 'longjohn', I obtained the following error message:
TypeError: Cannot read property 'resume' of null
at socketOnDrain (_http_server.js:442:20)
at ServerResponse.updateOutgoingData as _onPendingData
at ServerResponse.OutgoingMessage._send (_http_outgoing.js:128:14)
at ServerResponse.OutgoingMessage.write (_http_outgoing.js:470:16)
at ReadStream.ondata (_stream_readable.js:525:20)
at emitOne (events.js:77:13)
at ReadStream.emit (events.js:169:7)
at readableAddChunk (_stream_readable.js:146:16)
at ReadStream.Readable.push (_stream_readable.js:110:10)
at onread (fs.js:1744:12)
at FSReqWrap.wrapper as oncomplete
at ReadStream.Readable.on (_stream_readable.js:655:33)
at ReadStream.Readable.pipe (_stream_readable.js:522:7)
at SendStream.stream (C:\XXXX\node_modules\express\node_modules\send\lib\send.js:533:10)
at SendStream.send (C:\XXXX\node_modules\express\node_modules\send\lib\send.js:480:8)
at C:\XXXX\node_modules\express\node_modules\send\lib\send.js:392:10
at FSReqWrap.oncomplete (fs.js:82:15)
worker 1760 died
The problem only occurs on my production server (ovh, Windows Hyper-V, node v4.2.1) and not on my local development computer (Windows 10, node v4.2.0).
It is worthwhile mentioning that I am using socket.io. I am also using an external service called Global Counter that displays the number of online users. I am running setInterval() in one of the workers to remove expired records.
Could someone tell me the meaning of these error messages, how the error occurs, and how I should solve it?
The text was updated successfully, but these errors were encountered: