unhandled error event

Question

unhandled error event

saqebakhter opened this issue 9 years ago · comments

Had almost 2 days of uptime until this crash:

2015-5-19T21:52::23 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::23 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::23 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::25 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::25 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::26 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::26 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::26 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T21:52::26 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T21:52::28 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:52::28 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-19T21:56::48 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T21:56::48 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T21:57::26 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T21:57::26 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:1::48 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:1::48 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:2::7 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:2::7 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:2::7 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:2::7 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:3::49 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:3::49 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:5::4 - ^[[32minfo^[[39m: Starting to upload file Taken 2 (2012)-trailer.mov
2015-5-19T22:5::6 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:5::6 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-19T22:5::11 - ^[[31merror^[[39m: There was an error with getting a new resumable link
2015-5-19T22:5::11 - ^[[31merror^[[39m: errors=[domain=global, reason=authError, message=Invalid Credentials, locationType=header, location=Authorization], code=401, message=Invalid Credentials
events.js:141
throw er; // Unhandled 'error' event
^
Error: EMFILE: too many open files, open '/mnt/2TB1/cache/upload/6e0a025498f8dc2c47290578f69d7809'
at Error (native)

Eric Yen · Answer 1 · Wed May 20 2015 20:14:48 GMT+0800 (China Standard Time)

It should be fixed.
The other thing is, maybe you could try switching to io.js? It should work with all versions of iojs. (you would need to reinstall the npm modules.)

saqebakhter · Answer 2 · Wed May 20 2015 20:17:48 GMT+0800 (China Standard Time)

Hey

Thanks for fixing - I'm already using iojs - currently using 1.8.1 -
should I move to 2.x ?

thejinx0r mailto:notifications@github.com
May 20, 2015 at 8:14 AM

It should be fixed.
The other thing is, maybe you could try switching to io.js
http://iojs.org? It should work with all versions of iojs. (you
would need to reinstall the npm modules.)

—
Reply to this email directly or view it on GitHub
#57 (comment).

--Saqeb

Eric Yen · Answer 3 · Wed May 20 2015 20:25:57 GMT+0800 (China Standard Time)

Nevermind, I take it back, it's not fixed, I just got the same error.
And you can update to 2.x. I tested it earlier and I made sure it works. I'm running on it now.

saqebakhter · Answer 4 · Wed May 20 2015 20:32:03 GMT+0800 (China Standard Time)

:( thanks for trying ! I'm sure you'll get it haha.

thejinx0r mailto:notifications@github.com
May 20, 2015 at 8:25 AM

Nevermind, I take it back, it's not fixed, I just got the same error.

—
Reply to this email directly or view it on GitHub
#57 (comment).

--Saqeb

saqebakhter · Answer 5 · Wed May 20 2015 21:08:53 GMT+0800 (China Standard Time)

Okay I've switched to 2.x iojs.

saqebakhter · Answer 6 · Fri May 22 2015 04:07:45 GMT+0800 (China Standard Time)

On the latest commit 56c6744

Still getting this error:

2015-5-21T15:31::6 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-21T15:31::6 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-21T15:31::19 - ^[[31merror^[[39m: error after piping code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-21T15:31::19 - ^[[31merror^[[39m: code=ECONNRESET, errno=ECONNRESET, syscall=read
2015-5-21T15:31::34 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-21T15:31::34 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-21T15:31::35 - ^[[31merror^[[39m: error after piping code=EPIPE, errno=EPIPE, syscall=write, address=undefined
2015-5-21T15:31::35 - ^[[31merror^[[39m: code=EPIPE, errno=EPIPE, syscall=write, address=undefined
events.js:141
throw er; // Unhandled 'error' event
^
Error: EMFILE: too many open files, open '/mnt/2TB1/cache/upload/cfb1f6b51554fc1ec30f21a4a8359344'
at Error (native)

Eric Yen · Answer 7 · Fri May 22 2015 21:42:26 GMT+0800 (China Standard Time)

I've tracked down the error to file downloading.
There's a leak going on there with the streams.

saqebakhter · Answer 8 · Fri May 22 2015 21:51:05 GMT+0800 (China Standard Time)

Cool is the fix checked in ? I'll update to the latest and keep testing.

Eric Yen · Answer 9 · Sat May 23 2015 03:32:49 GMT+0800 (China Standard Time)

I don't have a proper fix yet. I keep getting closer.

Eric Yen · Answer 10 · Sat May 23 2015 18:34:43 GMT+0800 (China Standard Time)

This is fixed. I found the leak that caused the "EMFILE".

Eric Yen · Answer 11 · Sat May 23 2015 18:56:52 GMT+0800 (China Standard Time)

I've switched to a linux based os from OS X and the performance is great.
I'm now using 1mb chunks and 4 advanced chunks and it's surprisingly keeps up.

And on linux, the default setting for fuse is to run on multithread and on OS X, it's single threaded.
I think that's what was causing a lot of the problems that I wasn't getting before.

saqebakhter · Answer 12 · Sun May 24 2015 00:42:46 GMT+0800 (China Standard Time)

Could you share your config.json settings?

Eric Yen · Answer 13 · Sun May 24 2015 01:09:01 GMT+0800 (China Standard Time)

{
  "debug": true,
  "clientId": "520595891712-ofd914r711g7cuq8uofet8ekf7mneo2g.apps.googleusercontent.com",
  "clientSecret": "dzxWOFdSUtT7RL8bvrVMOK9t",
  "mountPoint": "/Users/eric/GDrive",
  "chunkSize": 1048576,
  "refreshDelay": 300000,
  "advancedChunks": 4,
  "maxCacheSize": 870400,
  "cacheLocation": "/GDrive/.gdrive",
  "maxConcurrentUploads": 4,
 }

Without the access token, this is my config file