Runner: Self-hosted runner failed to update itself

[arischow]

Describe the bug
So on 15 Jul we're still using a version of 2.263.0, and it tried to update itself to 2.267.1 yesterday but failed.

To Reproduce
Steps to reproduce the behavior:

1. Use 2.263.0 as self-hosted runner
2. And it will try to update itself to 2.267.1
3. When it run _update.sh, it failed.

Expected behavior
It would be updated successfully and process jobs properly.

Runner Version and Platform

2.263.0

OS of the machine running the runner?
We run our self-hosted runner as a container, which uses Ubuntu bionic, in our Kubernetes cluster.

What's not working?

The runner itself will crash.

Job Log Output

N/A

Runner and Worker's Diagnostic Logs

Runner update in progress, do not shutdown runner.
Downloading 2.267.1 runner
Downloading 2.267.1 runner
Downloading 2.267.1 runner
Waiting for current job finish running.
Generate and execute update script.
Runner will exit shortly for update, should back online within 10 seconds.
/tmp/_update.sh: line 31: ping: command not found
/tmp/_update.sh: line 31: ping: command not found

I think the problem is the update script assumes the OS would have ping binary, which is not reliable enough. IMHO it should run some pre-script before updating like install_dependencies.sh before starting the runner.

[gfvirga]

I just had this issue as well.
https://github.com/actions/runner/pull/647 seems to fix this.