You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Apr 22, 2023. It is now read-only.
I've got a CI system running on node v0.10.26 (+some in-house patches to aid in proxy debugging which I can't see causing any trouble in this case) which does a lot of child_process.spawn() calls, and I'm getting hangs where none of the child events get fired but the spawned process is a zombie. This is only happening so far on debian-7 instances running on google compute engine (Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.54-2 x86_64), and it only seems to be happening on short-lived processes (e.g. ln, git fetch when a repo is up-to-date).
This could obviously be any number of things in my patched version of node and the cloud instance I'm running on, but I was hoping someone had run into something similar and could give me some pointers. Am I right in assuming node should always be getting the SIGCHLD and firing some kind of event rather than sitting idle with a zombie? Is there a way to turn on low-level debugging for signal handling and/or uv_spawn? Thanks for any help...
The text was updated successfully, but these errors were encountered:
It seems like the SIGCHLD does get through eventually, but only after a long time (many minutes) of the child process sitting around as a zombie. Weird. Probably a GCE problem.
I've got a CI system running on node v0.10.26 (+some in-house patches to aid in proxy debugging which I can't see causing any trouble in this case) which does a lot of
child_process.spawn()
calls, and I'm getting hangs where none of the child events get fired but the spawned process is a zombie. This is only happening so far on debian-7 instances running on google compute engine (Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.54-2 x86_64), and it only seems to be happening on short-lived processes (e.g.ln
,git fetch
when a repo is up-to-date).This could obviously be any number of things in my patched version of node and the cloud instance I'm running on, but I was hoping someone had run into something similar and could give me some pointers. Am I right in assuming node should always be getting the
SIGCHLD
and firing some kind of event rather than sitting idle with a zombie? Is there a way to turn on low-level debugging for signal handling and/or uv_spawn? Thanks for any help...The text was updated successfully, but these errors were encountered: