| From | Sent On | Attachments |
|---|
| Subject: | [courier-users] Re: 432 Service temporarily unavailable | |
|---|---|---|
| From: | Sam Varshavchik (mrs...@courier-mta.com) | |
| Date: | Jan 25, 2002 3:29:40 pm | |
| List: | net.sourceforge.lists.courier-users | |
Refine Search
| From | Sent On | Attachments |
|---|---|---|
| Josh Zerlan | Jan 23, 2002 3:14 pm | |
| Sam Varshavchik | Jan 23, 2002 3:29 pm | |
| Josh Zerlan | Jan 24, 2002 5:08 pm | |
| Sam Varshavchik | Jan 24, 2002 8:52 pm | |
| Josh Zerlan | Jan 25, 2002 7:18 am | |
| Sam Varshavchik | Jan 25, 2002 3:29 pm | |
| Tomas Fasth | Jan 26, 2002 8:48 am | |
| Sam Varshavchik | Jan 26, 2002 5:00 pm |
| Subject: | [courier-users] Re: 432 Service temporarily unavailable | |
|---|---|---|
| From: | Sam Varshavchik (mrs...@courier-mta.com) | |
| Date: | Jan 25, 2002 3:29:40 pm | |
| List: | net.sourceforge.lists.courier-users | |
Josh Zerlan writes:
At 22:52 1/24/2002, you wrote:
Josh Zerlan writes:
select(1, [0], NULL, NULL, {600, 0}) = 1 (in [0], left {566, 480000}) read(0, "RCPT TO:<te...@cm.pcslab.com>\r\n", 5120) = 30 write(5, "te...@cm.pcslab.com\t\t\n", 21) = -1 EPIPE (Broken pipe) --- SIGPIPE (Broken pipe) --- write(1, "432 Service temporarily unavaila"..., 38) = 38 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGCHLD, NULL, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 nanosleep({8, 0}, {8, 0}) = 0 time([1011920440]) = 1011920440 time([1011920440]) = 1011920440 select(1, [0], NULL, NULL, {600, 0} -=-=-=-=-=-=-=-=-=-=-=-=-=- As you can see here, it immediately dies with a broken pipe... I really have no idea what it's trying to do here. Is it trying to call a function that doesn't work/exist? What's going on?
It's the pipe to the submit child process, which terminated for some reason. You need to use the -f option to strace to also trace the child processes, and use that to determine how submit terminated.
That strace _IS_ with the -f option, that's why I'm confused.
No it's not. If you did use the -f option, each line of the trace would be prefixed with the process id.
-- Sam

