varnishncsa dies on HUP signal.

andan andan armdan20 at gmail.com
Tue Jun 28 11:29:07 CEST 2011


Hello.

I'm testing Varnish 3.0 on CentOS 5 (2.6.18-238.5.1.el5 x86_64) using
rpms from http://repo.varnish-cache.org/.

varnishncsa dies on logrotate, the logrotate script do a kill -HUP to
force refresh log files, this works fine on varnish 2.1.5 and previous
versions.

Debugging with strace:

root     24777  0.0  0.0  95272   356 ?        Ss   10:59   0:00
/usr/bin/varnishncsa -c -w /var/log/varnish/varnishncsa.log -D -P
/var/run/varnishncsa.pid

$> strace -p 24777

Process 24777 attached - interrupt to quit
restart_syscall(<... resuming interrupted call ...>) = 0
nanosleep({0, 50000000}, NULL)          = 0
nanosleep({0, 50000000}, NULL)          = 0
nanosleep({0, 50000000}, NULL)          = 0
nanosleep({0, 50000000}, NULL)          = 0
....
nanosleep({0, 50000000}, NULL)          = 0
nanosleep({0, 50000000}, NULL)          = 0
nanosleep({0, 50000000}, 0)             = ? ERESTART_RESTARTBLOCK (To
be restarted)
--- SIGHUP (Hangup) @ 0 (0) ---
rt_sigreturn(0x1)                       = -1 EINTR (Interrupted system call)
write(2, "Assert error in vsl_nextlog(), v"..., 95) = 95
write(2, "  errno = 4 (Interrupted system "..., 38) = 38
rt_sigprocmask(SIG_UNBLOCK, [ABRT], NULL, 8) = 0
tgkill(24777, 24777, SIGABRT)           = 0
--- SIGABRT (Aborted) @ 0 (0) ---
Process 24777 detached

Thanks in advance.
Kind Regards.




More information about the varnish-misc mailing list