When Varnish restarts the SHM log file gets abandoned and Varnishkafka exits gracefully as consequence. It would be nice to implement a "auto-reconnect" behavior like varnishlog:
"""
$ sudo varnishlog -g request
Log abandoned
Log reacquired
"""
We are currently rely on systemd to restart Varnishkafka when it fails or shutsdown.
Useful snippet from VUT:
https://github.com/varnishcache/varnish-cache/blob/4.1/lib/libvarnishtools/vut.c#L366
Steps for this task:
- Add the code necessary to implement the re-connect feature to Varnishkafka
- Code Review, tests
- Debian packaging
- Deploy to cp Misc and Maps
- Deploy to Upload and Text