Date: Sun, 11 Apr 2010 12:12:10 +0300 From: Mikolaj Golub <to.my.trociny@gmail.com> To: freebsd-fs@FreeBSD.org Subject: hastd: socket leakage on worker exit Message-ID: <86pr265p5h.fsf@kopusha.onet>
next in thread | raw e-mail | index | archive | help
--=-=-=
Hi,
Playing with HAST I have noticed the following issue with hasd:
....
Apr 10 22:32:36 hastb hastd: [storage] (secondary) Split-brain detected, exiting.
Apr 10 22:32:36 hastb hastd: [storage] (secondary) Worker process failed (pid=6474, status=78).
Apr 10 22:32:56 hastb hastd: [storage] (secondary) Split-brain detected, exiting.
Apr 10 22:32:56 hastb hastd: [storage] (secondary) Worker process failed (pid=6475, status=78).
Apr 10 22:32:56 hastb hastd: [storage] (secondary) Unable to create control sockets be: Too many open files
And sockstat:
root hastd 711 4 stream /var/run/hastctl
root hastd 711 5 tcp4 *:8457 *:*
root hastd 711 7 dgram -> /var/run/logpriv
root hastd 711 8 stream (not connected)
root hastd 711 9 stream -> ??
root hastd 711 10 stream -> ??
root hastd 711 12 stream -> ??
root hastd 711 13 stream -> ??
root hastd 711 14 stream -> ??
root hastd 711 15 stream -> ??
[ ... and so on .. ]
The patch below has fixed the issue.
--
Mikolaj Golub
--=-=-=
Content-Type: text/x-diff
Content-Disposition: inline; filename=hastd.c.close_on_child_exit.patch
--- sbin/hastd/hastd.c.orig 2010-04-11 11:52:10.000000000 +0300
+++ sbin/hastd/hastd.c 2010-04-11 11:51:23.000000000 +0300
@@ -138,6 +138,7 @@ child_exit(void)
(unsigned int)pid, WEXITSTATUS(status));
}
res->hr_workerpid = 0;
+ proto_close(res->hr_ctrl);
if (res->hr_role == HAST_ROLE_PRIMARY) {
sleep(1);
pjdlog_info("Restarting worker process.");
--=-=-=--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86pr265p5h.fsf>