Skip to content

GitLab

  • Menu
Projects Groups Snippets
    • Loading...
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
    • Contribute to GitLab
  • Sign in
  • G gitlabhq1
  • Project information
    • Project information
    • Activity
    • Labels
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributors
    • Graph
    • Compare
  • Issues 21
    • Issues 21
    • List
    • Boards
    • Service Desk
    • Milestones
  • Merge requests 12
    • Merge requests 12
  • CI/CD
    • CI/CD
    • Pipelines
    • Jobs
    • Schedules
  • Deployments
    • Deployments
    • Environments
    • Releases
  • Monitor
    • Monitor
    • Incidents
  • Packages & Registries
    • Packages & Registries
    • Package Registry
    • Infrastructure Registry
  • Analytics
    • Analytics
    • CI/CD
    • Repository
    • Value stream
  • Wiki
    • Wiki
  • Snippets
    • Snippets
  • Activity
  • Graph
  • Create a new issue
  • Jobs
  • Commits
  • Issue Boards
Collapse sidebar
  • gpt
  • large_projects
  • gitlabhq1
  • Issues
  • #7374

Closed
Open
Created Jul 25, 2014 by Administrator@rootOwner

Unicorn regularly times out

Created by: hugelgupf

Loading any GitLab page always takes 30+ seconds. During that time, one CPU core is at 100%.

Specs: 3 cores 2.1Ghz each, 2GB RAM; one empty repo in GitLab so far. Using gitlab 7.1.0, ruby 2.1.1. The gitlab check rake shows no problems.

This is what the unicorn log tells me every once in a while

E, [2014-07-25T23:01:28.563369 #2023] ERROR -- : worker=1 PID:2343 timeout (31s > 30s), killing
E, [2014-07-25T23:01:28.583393 #2023] ERROR -- : reaped #<Process::Status: pid 2343 SIGKILL (signal 9)> worker=1
I, [2014-07-25T23:01:28.588358 #2422]  INFO -- : worker=1 ready

production.log -- look at those time stamps!

Started GET "/git/cse113/hw1" for 129.138.6.215 at 2014-07-25 23:00:57 +0000
Processing by ProjectsController#show as HTML
  Parameters: {"id"=>"cse113/hw1"}
Completed 200 OK in 214ms (Views: 177.0ms | ActiveRecord: 5.8ms)
Started GET "/git/assets/application-8afcf791d7afe64a25e5240b75fcf084.css" for 129.138.6.215 at 2014-07-25 23:00:57 +0000
Started GET "/git/assets/application-75808322fdee0d74a99d5ed16b5493ca.js" for 129.138.6.215 at 2014-07-25 23:00:57 +0000
Started GET "/git/assets/print-1df3ea9b8ff148a6745321899e0cb213.css" for 129.138.6.215 at 2014-07-25 23:00:58 +0000
Started GET "/git/assets/logo-white-00839569c4e0e949538f31353fa20693.png" for 129.138.6.215 at 2014-07-25 23:01:28 +0000
Started GET "/git/assets/fontawesome-webfont-1762ab9ef23ec44fbaab6d1360c48988.woff" for 129.138.6.215 at 2014-07-25 23:01:29 +0000
Started GET "/git/assets/icon-search-d671d72e39d229479d0a5575acc802bb.png" for 129.138.6.215 at 2014-07-25 23:01:29 +0000

There are exactly 30 seconds between print and logo-white, suggesting some kind of timeout. The time of the logo-white request coincides with unicorn's timeout. Every request is this way -- application-(.).(js|css) and print-(.).css are requested, then a 30 second timeout, then the rest.

Any ideas?

Assignee
Assign to
Time tracking