You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
az throws an error when you've already got the cluster creds. This is masked by CI but breaks when running on your own device.
$ hubploy deploy data100 hub staging
CloudName IsDefault Name State TenantId
----------- ----------- ----------- ------- ------------------------------------
AzureCloud True data100-19s Enabled abcde....
A different object named data100-fall-2019 already exists in contexts
Traceback (most recent call last):
File "/home/ryan/.local/bin/hubploy", line 10, in <module>
sys.exit(main())
File "/home/ryan/.local/lib/python3.7/site-packages/hubploy/__main__.py", line 64, in main
auth.cluster_auth(args.deployment)
File "/home/ryan/.local/lib/python3.7/site-packages/hubploy/auth.py", line 150, in cluster_auth
deployment, **cluster['azure']
File "/home/ryan/.local/lib/python3.7/site-packages/hubploy/auth.py", line 227, in cluster_auth_azure
'--resource-group', resource_group
File "/usr/local/linux/anaconda3.7/lib/python3.7/subprocess.py", line 328, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['az', 'aks', 'get-credentials', '--name', 'data100-fall-2019', '--resource-group', 'data100-fall-2019']' returned non-zero exit status 1.
$ az aks get-credentials -n data100-fall-2019 -g data100-fall-2019
A different object named data100-fall-2019 already exists in contexts
The text was updated successfully, but these errors were encountered:
Sorry I missed this question @ryanlovett. Did you solve it? Because a CI should start fresh, I don't think it should have the cluster in its context. Although maybe it could with caching. Locally you can use --overwrite-existing to overwrite (and with the newest versions of the Azure CLI, I think it defaults to asking). We could consider adding this flag to the Azure auth function if you think that is a good idea.
az throws an error when you've already got the cluster creds. This is masked by CI but breaks when running on your own device.
The text was updated successfully, but these errors were encountered: