-
Notifications
You must be signed in to change notification settings - Fork 3.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Script disconnected unexpectedly while long curl #7323
Comments
I tried today to use |
Any update on this? I am also facing the same issue if a command runs for longer time packer build process is killed. |
@umerebryx |
Hello there; the original gist posted is a 404. But I think this issue could be due to the fact that the communicator will hang up after a while if nothing happens. |
@azr In my case the curl was against docker domain (getting the installation file) |
Do you have the configuration to reproduce it ? |
In my case it abruptly fails while the script is running and there is continuous output. Here is where my build stops:
Another thing I observed is the script always fails at 6 mins and ~47 sec which make me think if there is a global timeout set somewhere that is causing this behaviour? My packer version is 1.4.3 and running it on Mac. |
Do you have easy to reproduce steps ? It looks like to me that the command crashed:
OOM ? |
I will setup a temp repo for you to reproduce this. I have done another test by running same build script from an ec2 instance c4.2xlarge and this time everything completed without any issues. I will double check my local setup to see this is not a memory issue as you mentioned (but shouldn't OOM through a segmentation fault locally?). |
This is happening for me every other build or so, it's pretty inconsistent.
|
@anthonymag , for me the issue was fix after I freed up some memory. The worst part is packer doesn't give any indication the termination is because of OOM. You can try running by closing all your programs to see if the issue is RAM in your case. Hope this helps. |
@umerebryx thanks for your response! I'm surprised as I'm giving the built instance 8GB, but I can increase this and try it out, unless you mean memory on the Packer host that is remotely controlling the build? |
You have to free up memory on the host that is running the packer build command and not the remote host. |
An interesting twist in the story... |
I'm currently running the build directly from a Jenkins host, and will change it to run from a Jenkins worker/slave to have its own dedicated resources and will let you guys know if that fixes the issue. It might have failed from OOM whenever it was running alongside another build. |
No dice. I gave it its own dedicated host with 8GB and I still get the same failure every other build or so... |
also got similar issue with long running "shell" provisionner that perform a lot of network i/o (conda) that end up mid-way with:
without the actual error being logged. running |
@yaniveliash, I have been trying to use all the suggested reproduction steps mentioned in the discussion, however, I cannot reproduce this issue. Maybe it was just a memory issue as indicated by a user. |
Seems like long curl (about 2 minutes)
disconnect my session in the middle.
That isn't consist as sometime it works and sometimes not...
Packer version : 1.3.3
Platform : AWS EBS (Ubuntu)
Debug log:
https://gist.github.com/geostant/16f8ade96ecbdecc0a30e99df6ad7135
The text was updated successfully, but these errors were encountered: