Because of differences between Cloud Foundry API (CAPI) v2 and v3, cf connect-to-service fails on applications that have been deployed with a zero downtime push, using cf CLI v6.
If you have deployed an app with cf v3-zdt-push, you have to use cf v3-ssh or the cf CLI v7. cf ssh under cf CLI v6 will fail.
Pushing updates to your app with a v3-zdt command causes the new web process and app GUID to mismatch. cf ssh does not handle this scenario. You must use the cf v3-ssh command instead.
Note: This limitation only applies to cf CLI v6.
Because of differences between Cloud Foundry API (CAPI) v2 and v3,
cf connect-to-service
fails on applications that have been deployed with a zero downtime push, using cf CLI v6.If you have deployed an app with
cf v3-zdt-push
, you have to usecf v3-ssh
or the cf CLI v7.cf ssh
under cf CLI v6 will fail.From Rolling App Deployments (Beta):
Steps to reproduce:
cf v3-zdt-push
cf connect-to-service APP-NAME SERVICE-NAME
. That will result in:Executing
cf ssh APP-NAME
will result in the same error.Workaround is to use
cf v3-ssh
orcf7 ssh
to do local port forwarding, and then use the service client (e.g.,psql
) to connect to that local port.