aws-vault server sticks around longer than necessary

[mtibben]

When using aws-vault exec with the --server flag, the server process sticks around even once the subshell has gone away.

$ aws-vault exec --server contests
$ exit
$ ps a | grep aws-vault
46031 s000  S      0:00.00 sudo -b aws-vault server
46032 s000  S      0:00.04 aws-vault server
46173 s000  R+     0:00.00 grep aws-vault

[lox]

That is by design. The server is a dumb proxy that listens on the metadata address and forwards requests to the ephemeral server that runs in the aws-vault exec process.

[mtibben]

I think though that aws-vault (as a good parent process) should be responsible for stopping it when it's no longer required, or alternatively we should be giving control of it to launchctl

[lox]

The issue is it needs root access to launch. I wanted to avoid prompting for root creds constantly.

Thoughts @pda?

On 11 Dec 2017, at 8:49 am, Michael Tibben [e-mail] wrote:

I think though that aws-vault (as a good parent process) should be responsible for stopping it when it's no longer required, or alternatively we should be giving control of it to launchctl

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

[pecigonzalo]

I agree about the constant promts, maybe we can have
A) a parameter (remain or not)
B) a way to terminate it like aws-vault server stop and start or similar (start is automatically done if aws-vault exec --server $profile

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[tlaukkan]

Any progress on this one?

[daneah]

I find that occasionally the process that sticks around (it's aws-vault proxy, these days) contributes heavily to the CPU load of my machine—this is typically when aws-vault exits of its own accord rather than when I exit the subshell explicitly, though. That's perhaps a separate issue I could report, but I would agree that I normally expect a process that creates a subshell to clean itself up when that subshell is exited.