We're using collectd instance to translate from collectd protocol to riemann so config (in attachment) is pretty simple; with enabled write_riemann plugin we are getting segfault as soon as it starts receiving data: /usr/sbin/collectd -C /etc/collectd-server.conf -f [2024-11-27 16:57:34] plugin_load: plugin "logfile" successfully loaded. Segmentation fault (as in it starts when I block incoming traffic in iptables but crashes when I remove the block and traffic hits it). weirdly enough, not on every server, but I assume that's coz traffic is hashed and some get the type of packets that trigger that bug. I think it might be this one https://github.com/collectd/collectd/issues/4050 that was patched and backported to 5.12 at beginning of this year: https://github.com/collectd/collectd/pull/4232 https://github.com/collectd/collectd/commit/cfdf526f1b4867d8952fc6d37b85f0d426a78fc7
Hi Mariusz, Just had a look at https://github.com/collectd/collectd/compare/collectd-5.12.0...collectd-5.12 That diff is not even small, and there was no release since 5.12.0.... Guess I should just start to pick changes, no clue if 5.13 will ever be released as the work on 6.0 is also "progressing" since several years now... I'll have a look into merging alls those changes. Bernd
Hi Bernd Yeah, looking at it I wouldn't expect anything in 5.x branch sans for bugfixes, as even in 6.x development is going slowly. Kinda not surprised by low traffic as 6 is basically "break everything" change and if someone wanted 6.x to have better impedance matching with Prometheus there are currenly more mature prom-compatible projects. Weirdly enough we were on that version since basically current stable and had zero problem till few days ago it just started segfaulting on that plugin Cheers Mariusz