refed from nodejs/build: CITGM runs for 16 hours starving other jobs

[refack]

• Node Version: master
• CitGM Version:
• Platform: Our CI

How can I help?

For example this one has been waiting for a fedore22 worker for 6 hours

Ref: nodejs/build#719

[gibfahn]

So I ran on v7.10.0 and it took 90 minutes

The next build ran on nodejs/node#13011 and was still hanging on fedora22 after 7 hours.

Last bit of output was this, but as we're running in parallel it's hard to say whether this module is the one that has issues.

verbose: esprima using-uid   | 1000                
verbose: esprima using-gid   | 1000                
verbose: esprima using-node  | /home/iojs/build/workspace/citgm-smoker/nodes/fedora22/smoker/bin/node
verbose: esprima using-npm   | /home/iojs/build/workspace/citgm-smoker/nodes/fedora22/smoker/bin/npm
verbose: esprima mk.tempdir  | /home/iojs/build/workspace/citgm-smoker/nodes/fedora22/citgm_tmp/2648f141-f522-42eb-be7d-1a1cda3d370c
info: lookup              | esprima             
info: lookup-found        | esprima             
info: esprima lookup-replace| https://github.com/jquery/esprima/archive/3.1.3.tar.gz
info: esprima npm:        | Downloading project: https://github.com/jquery/esprima/archive/3.1.3.tar.gz
info: esprima npm:        | Project downloaded esprima-3.1.3.tgz
info: esprima npm:        | npm install started 
verbose: esprima npm:        | Using temp directory: "/home/iojs/build/workspace/citgm-smoker/nodes/fedora22/citgm_tmp/2648f141-f522-42eb-be7d-1a1cda3d370c/npm_config_tmp"
verbose: esprima nodedir     | Using --nodedir="/home/iojs/build/workspace/citgm-smoker/nodes/fedora22/node"
error: acorn npm:          | npm-install Timed Out
error: failure             | Install Timed Out   
info: duration            | test duration: 602913ms
error: acorn done          | done - the test suite for acorn version 5.0.3 failed
info: esprima npm:        | npm install successfully completed
info: esprima npm:        | test suite started  

@refack is it just fedora22 that seems to be having the problem?

[refack]

@refack is it just fedora22 that seems to be having the problem?

It's just that fedora22 is has a single instance so it popped up as starving everything else. I think I saw osx1010 also starved.

[MylesBorins]

We have a time out mechanism in citgm… very odd that it is not working

…

On May 13, 2017, at 1:28 PM, Refael Ackermann ***@***.***> wrote: @refack <https://github.com/refack> is it just fedora22 that seems to be having the problem? It's just that fedora22 is has a single instance so it popped up as starving everything else. I think I saw osx1010 also starved. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#409 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAecVyozbS2HAGFqcw5DZWTCGNJse4g-ks5r5egwgaJpZM4NaGma>.

[targos]

Anything actionable on this?

[refack]

I believe this was mitigated with #441