Bug#832637: More info

[Brian Potkin <claremont102@gmail.com>]

On Sun 21 Aug 2016 at 17:30:10 +0200, Roderich Schupp wrote:

> On Sun, Aug 21, 2016 at 4:27 PM, Brian Potkin <claremont102@gmail.com>
> wrote:
> 
> > Are you positive cups.service is still running while cups-browsed is
> > stopping? 'systemctl status cups' in another terminal would tell you.
> >
> 
> Yes
> 
> 
> > If I use the sequence
> >
> >   systemctl start cups cups-browsed.service
> >   systemctl stop cups-browsed.service
> >
> > it stops cups-browsed immediately here.
> >
> 
> I just did:
> 
> # systemctl start cups cups-browsed.service
> # systemctl status cups cups-browsed.service
> cups.service - CUPS Scheduler
>    Loaded: loaded (/lib/systemd/system/cups.service; enabled; vendor
> preset: ena
>    Active: active (running) since Sun 2016-08-21 09:14:34 CEST; 8h ago
>    ...
> cups-browsed.service - Make remote CUPS printers available locally
>    Loaded: loaded (/lib/systemd/system/cups-browsed.service; enabled;
> vendor pre
>    Active: active (running) since Sun 2016-08-21 17:16:33 CEST; 9s ago
>    ...
> # systemctl stop cups-browsed.service
> (waits 90 sec)
> 
> If I strace the cups-browsed process (all threads) during tese 90 seconds I
> see it repeats the following until it gets killed:
> 
> 19552 connect(10, {sa_family=AF_INET6, sin6_port=htons(631),
> inet_pton(AF_INET6, "::1", &sin6_addr), sin6_flowinfo=0, sin6_scope_id=0},
> 28) = -1 EINPROGRESS (Operation now in progress)
> 19552 fcntl(10, F_SETFL, O_RDWR)        = 0
> 19552 poll([{fd=10, events=POLLIN|POLLOUT}], 1, 100) = 1 ([{fd=10,
> revents=POLLIN|POLLOUT|POLLERR|POLLHUP}])
> 19552 sendto(10, "POST / HTTP/1.1\r\nContent-Length:"..., 169, 0, NULL, 0)
> = -1 ECONNREFUSED (Connection refused)
> 19552 sendto(10, "POST / HTTP/1.1\r\nContent-Length:"..., 169, 0, NULL, 0)
> = -1 EPIPE (Broken pipe)
> 19552 --- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=19552,
> si_uid=0} ---
> 19552 sendto(10, "POST / HTTP/1.1\r\nContent-Length:"..., 169, 0, NULL, 0)
> = -1 EPIPE (Broken pipe)
> 19552 --- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=19552,
> si_uid=0} ---
> 19552 close(10)                         = 0
> 19552 socket(AF_INET6, SOCK_STREAM, IPPROTO_IP) = 10
> 19552 setsockopt(10, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
> 19552 setsockopt(10, SOL_SOCKET, SO_REUSEPORT, [1], 4) = 0
> 19552 setsockopt(10, SOL_TCP, TCP_NODELAY, [1], 4) = 0
> 19552 fcntl(10, F_SETFD, FD_CLOEXEC)    = 0
> 19552 fcntl(10, F_GETFL)                = 0x2 (flags O_RDWR)
> 19552 fcntl(10, F_SETFL, O_RDWR|O_NONBLOCK) = 0
> 19552 connect(10, {sa_family=AF_INET6, sin6_port=htons(631),
> inet_pton(AF_INET6, "::1", &sin6_addr), sin6_flowinfo=0, sin6_scope_id=0},
> 28) = -1 EINPROGRESS (Operation now in progress)
> 19552 fcntl(10, F_SETFL, O_RDWR)        = 0
> 19552 poll([{fd=10, events=POLLIN|POLLOUT}], 1, 100) = 1 ([{fd=10,
> revents=POLLIN|POLLOUT|POLLERR|POLLHUP}])
> 19552 sendto(10, "POST / HTTP/1.1\r\nContent-Length:"..., 169, 0, NULL, 0)
> = -1 ECONNREFUSED (Connection refused)
> ...

I have no special skills in interpreting strace outputs but this would
appear to be confirmation that cups-browsed cannot contact cups when it
closes down. But cups is running, so it is beginning to pass through my
mind that your issue is not the same as Heinrich's.

How do you go on with

  systemctl start cups.service cups-browsed.service
  systemctl stop cups.service
  systemctl status cups-browsed.service

Is there a 90 second hang when cups is stopped?

Cheers,

Brian.