Node Process crashed/aborted by SIG6

[rohit12sh]

Node process is getting killed/aborted by server after 30-40min of running:

OS: Red Hat Enterprise Linux Server release 6.5
Arch: x86_64
Kernel: 2.6.32-431.20.3.el6.x86_64

Process /bin/node was killed by signal 6 (SIGABRT)
Program terminated with signal 6, Aborted.
#0 0x000000396c00f5db in raise () from /lib64/libpthread.so.0

Not sure where and why the program is getting aborted by SIG6. I am using strong-oracle module and doing long running data migration.

[mscdex]

Are you able to reproduce it without using third party modules? Also what node version is this?

[rohit12sh]

Haven't seen this issue before with Nodejs. Seeing it since using third-party module. I understand it could be nothing to do with Nodejs, but wanted to confirm it.

Node version: v0.10.36

[cjihrig]

Does the same thing happen with a newer version of Node by any chance?

[rohit12sh]

@cjihrig:
strong-oracle module doesn't work with latest version of Node - v4.x. Any other version you would recommend to try?

[cjihrig]

Sorry, I don't really know enough about strong-oracle. Does it work with 0.12?

[mscdex]

@rohit12sh As a general debugging thing, you might try executing your script using the debug version of node under gdb. When it crashes, get a backtrace and that may help pinpoint the source of the error.

[rohit12sh]

@cjihrig : getting 'Module did not self-register.' for Strong-Oracle when trying to run process with node-v0.12.7-linux-x64

[mscdex]

@rohit12sh You have to rebuild/reinstall the module after you upgrade between "major versions" like 0.10, 0.12, and 4.0 because of C++ API changes. strong-oracle seems to be using nan though, so it should work for all of those versions after rebuilding/reinstalling.

[rohit12sh]

I have rebuilt the module, npm rebuild, rm -rf node_modules/. I keep getting Error: Module did not self-register.

I ran these:
node-v0.12.7-linux-x64/bin/npm install
node-v0.12.7-linux-x64/bin/node script.js

any ideas?

[cjihrig]

When you say rm -rf node_modules/ do you mean you wiped out the directory and ran npm install again?

[rohit12sh]

@cjihrig
yes. wiped and reinstalled.

one thing I tried was running node with more memory. This time it didn't crash.
node --max-old-space-size=6000

[seishun]

Probably strong-oracle is leaking memory? You should probably submit your issue there.

[rohit12sh]

@seishun: That's what I am coming down to. I have submitted the issue there already. Currently using memwatch to collect some stats on memory leak.

For now I am tweaking my program to not load large data in memory or increase memory of Node when running the process (node -max-old-space-size=4096).

Can some confirm if it true that I can't have 4G heap memory assigned to Node on 64-bit. Does it change with Node v4.x?

[bnoordhuis]

Closing, inactive for > 1 year.

Can some confirm if it true that I can't have 4G heap memory assigned to Node on 64-bit.

That was a bug in older versions of V8, yes, but has since been fixed.