It was fixed in rspamd somewhere between 1.9.4 and 17 Sept (which covers
some quite major changes as it switched to a different event loop library),
but somewhere between then and 6 Oct it started segfaulting at startup in
rspamd_mempool_new_ which I haven't tracked down yet.
--
Sent from a phone, apologies for poor formatting.
On 28 October 2019 04:33:32 Jordan Geoghegan <[email protected]> wrote:
On 2019-10-27 17:29, Chris Narkiewicz wrote:
Rspamd stop rc script doesn't work in OpenBSD 6.6.
1. Fresh OpenBSD 6.6 installation
2. pkg_add rspamd
3. rcctl start rspamd
Works.
4. rcctl stop rspamd timeouts
Looking at rspamd logs, it looks like it doesn not work
well with SIGTERM. It waits for workers.
Currently I work around it by adding custom rc_stop():
rc_stop() {
pkill -KILL -T "${daemon_rtable}" -xf "${pexp}"
pkill -KILL -u _rspamd
}
Anybody can confirm if this is a problem?
Woops, meant to send this to ports@....
-------------------------
Yes, can confirm. I think I read something the other day mentioning
sthen@ possibly having a diff floating around to fix the issue.
Hopefully the fix will be backported to -stable.
