0

Recently I was trying to do a blue green deployment of my marathon app by using the marathon-lb zdd.py script as mentioned in marathon-lb github documentation (https://github.com/mesosphere/marathon-lb#zero-downtime-deployments). And I had the issue while telling the script to deploy the current version(green one) completely, it doesn't kill and scale to zero the last single task of the blue or old deployment. So my blue app always remains at 1. Is there anywhere a better documentation of zdd.py rather than the marathon-lb github documentation?

Tinkaal Gogoi
  • 4,344
  • 4
  • 27
  • 36
  • 1
    Can you provide a little more information? Perhaps attach the output from `zdd.py`? What type of application are you running? What's the target # of instances? – Brenden Matthews Oct 31 '16 at 12:32
  • `2016-10-28 12:41:04,512 zdd: Coverting all instances to current config 2016-10-28 12:41:04,512 zdd: Current config color is green 2016-10-28 12:41:04,512 zdd: Considering blue color as existing app and green color as new app 2016-10-28 12:41:09,525 zdd: Existing app running 2 instances, new app running 2 instances 2016-10-28 12:41:09,533 zdd: Found 4 app listeners across 1 HAProxy instances 2016-10-28 12:46:05,236 zdd: Timed out waiting for tasks to fully drain, find any draining tasks and continue with deployment... 2016-10-28 12:46:05,236 zdd: Scaling new app up to 4 instances` – Tinkaal Gogoi Oct 31 '16 at 15:08
  • `Existing app running 2 instances, new app running 4 instances 2016-10-28 12:46:15,307 zdd: Found 6 app listeners across 1 HAProxy instances ... 2016-10-28 12:51:11,009 zdd: Timed out waiting for tasks to fully drain, find any draining tasks and continue with deployment... 2016-10-28 12:51:11,009 zdd: There are 1 draining listeners, about to kill the following tasks: - directory-service-blue.3736d0db-9d07-11e6-b59a-000c29960c71 2016-10-28 12:51:11,009 zdd: Scaling down old app by 1 instances 2016-10-28 12:51:16,049 zdd: Existing app running 1 instances, new app running 4 instances` – Tinkaal Gogoi Oct 31 '16 at 15:11
  • `2016-10-28 12:51:21,062 zdd: Existing app running 1 instances, new app running 4 instances 2016-10-28 12:51:21,070 zdd: Found 5 app listeners across 1 HAProxy instances` – Tinkaal Gogoi Oct 31 '16 at 15:12
  • I am running a simple service in docker containers with 4 target instances . So the the last two lines keeps looping forever – Tinkaal Gogoi Oct 31 '16 at 15:12
  • Odd, looks like it's timing out but possibly failing at the last step. Let me test it and get back to you. – Brenden Matthews Nov 01 '16 at 18:25

0 Answers0