Uploaded image for project: 'National Data Service'
  1. National Data Service
  2. NDS-998

Intermittent etcd timeouts from apiserver

XMLWordPrintableJSON

    • Icon: Bug Bug
    • Resolution: Fixed
    • Icon: Normal Normal
    • None
    • None
    • Backend, User Interface
    • None
    • NDS Sprint 42, NDS Sprint 43

      When attempting to sign-up via the UI, etcd threw an error. Clicking the Submit button again threw an error: "Username is already in use".

      Relevant log output:

      E0809 15:58:14.812512      23 etcd.go:113] client: etcd cluster is unavailable or misconfigured; error #0: client: endpoint http://192.168.100.171:2379 exceeded header timeout
      		 
      E0809 15:58:14.812594      23 server.go:743] client: etcd cluster is unavailable or misconfigured; error #0: client: endpoint http://192.168.100.171:2379 exceeded header timeout
      		 
      09/Aug/2017:15:58:13 +0000 500 1004638μs "POST /api/register HTTP/1.1" - "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36"
      		 
      I0809 15:58:35.353894      23 server.go:256] remoteAddr: 10.100.7.1:49536
      		 
      09/Aug/2017:15:58:35 +0000 409 2472μs "POST /api/register HTTP/1.1" - "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36"

      We should look into one or both of the following:

      • Discover why these etcd failures occur and handle them gracefully within the API server (i.e. don't throw an error)
      • Determine how to detect this error in the UI and ignore it appropriately?

              lambert8 Sara Lambert
              lambert8 Sara Lambert
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: