Running multiple instances of skaffold no longer possible

[pre]

Expected behavior

Have two projects with A/skaffold.yml and B/skaffold.yml.

Expected: running skaffold dev in both directories A and B would work. This used to work in an earlier version still in January 2019.

Actual behavior

Running skaffold dev for project A will work as expected, but skaffold dev for project B will fail.

The error message is:

FATA[0000] creating runner: creating runner: initializing skaffold event handler: creating status server: creating listener: listen tcp :50051: bind: address already in use

Information

• Skaffold version: fails with v0.24.0
• Operating system: MacOS Mojave
• Contents of skaffold.yaml: anything

Steps to reproduce the behavior

1. Create A/skaffold.yml
2. Create B/skaffold.yml
3. Run skaffold dev under A
4. Run skaffold dev under B
   -> fails with listen tcp :50051: bind: address already in use (this used to work still in January 2019)

[matti]

skaffold dev --rpc-port :50052

seems to work.

[tejal29]

cc @nkubala who might have a better idea since this seems relevant to the Event API work he did.

[nkubala]

@pre thanks for the issue. I opened a PR that will fix this, but in the meantime you can use @matti's workaround to keep running skaffold.

[tejal29]

@nkubala Can you point us to the workaround? is it same as using --rpc-port flag?

[nkubala]

@tejal29 yeah specifying another rpc port fixes it. the previous logic to choose a new port if the provided/default one is already in use was busted.