surface hook errors
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Expired
|
High
|
Unassigned |
Bug Description
[Edit: Juju was working fine, except that it was swallowing an error from my hook making it difficult to diagnose and look like Juju was at fault]
I just bootstrapped a fresh lxd controller and deployed cs:postgresql and cs:ntp to the default model, and related postgresql and ntp.
The ntp install hook failed with:
unit-ntp-0: 16:45:55 ERROR juju.worker.
Automatic retries also fail:
unit-ntp-0: 16:45:56 INFO juju.worker.uniter found queued "install" hook
unit-ntp-0: 16:45:56 ERROR juju.worker.
That last error is peculiar, because the file actually exists:
ubuntu@
lrwxrwxrwx 1 root root 12 Oct 3 09:45 /var/lib/
ubuntu@
-rwxr-xr-x 1 root root 5019 Oct 3 09:45 ntp_hooks.py
ubuntu@
ubuntu@
description: | updated |
tags: | added: logging |
Changed in juju: | |
status: | New → Triaged |
importance: | Undecided → Medium |
milestone: | none → 2.1.0 |
tags: | added: metrics-collector |
In debug-hooks:
root@juju- 517c57- 0:/var/ lib/juju/ agents/ unit-ntp- 0/charm# hooks/install
bash: hooks/install: /usr/bin/python: bad interpreter: No such file or directory
Juju needs to log this failure, rather than hide it with a message that makes it look like Juju's fault.