Failed to start but no error about ansible-role-varnish

[geerlingguy]

Which OS was this occurring on, and with what version of Varnish? Usually this indicates a problem with the service's init script, and in Varnish's case, that init script is installed as part of the package. If the init script always returns a 0 (success), then there might not be much Ansible or this role can do...

[swoodford]

I just spent 3 hours debugging this same issue using Amazon AMI Linux. The issue was the deprecated -w flag in DAEMON_OPTS under templates/varnish.j2, as soon as that flag was changed from

-w ${VARNISH_MIN_THREADS},${VARNISH_MAX_THREADS},${VARNISH_THREAD_TIMEOUT} \

to

-p thread_pool_min=${VARNISH_MIN_THREADS} \
-p thread_pool_max=${VARNISH_MAX_THREADS} \
-p thread_pool_timeout=${VARNISH_THREAD_TIMEOUT} \

it was working.

Also FYI the Ansible Galaxy role has not been updated with the fixes from this GitHub repo. That would have saved me the 3 hours if the fixes were sync'd back to Ansible Galaxy.

[geerlingguy]

@swoodford - The latest tag in Galaxy is 1.3.0, which corresponds with the same tag here (commit 0a0a5a9). There's no newer code I can send to Galaxy at this point...

See also: #17 — were you installing 4.1, or 4.0?

[swoodford]

@geerlingguy Sorry, you are correct sir. After some research I discovered a colleague inserted that -w flag after copying an old Varnish 3 config.

But hopefully @CashWilliams has a similar problem which is causing the service to fail to start without a helpful error message or log output.

[geerlingguy]

@swoodford - Thanks for the update!

[geerlingguy]

So it's likely a config/VCL issue outside this role's purview—I'm going to mark this as closed, but feel free to follow up if any other issues are found!