Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 12 Jan 2006 08:46:14 +0100 (CET)
From: Bernhard Fischer <bernhard@...chli.org>
To: owl-users@...ts.openwall.com
Subject: Re: ntpd terminates right after starting

Hi Alexander,

this behavior occurs not on every machine from mine as i said before,
but always when i activate the Link Eqaulizer net/sched/sch_teql.c
(CONFIG_NET_SCH_TEQL not as module) which itself generates a new teql
network-device. So what?!? Does it also count that there are 4 eths on
the problematic one and only 2 eths on the other one? I will test this
machine with the same network components.
strace does not show as much as hoped for. It would be more important to
see when/why the forked client terminates. I have to do more logging!

Here are the last 30 lines of strace from "ntpd -d -f /etc/ntpd.conf":
....
read(3, "J\211\213\4^CQ}\2\222\263\301G\267%F4\373U\276\341iZ\244"...,
32) = 32
close(3)                                = 0
getpid()                                = 408
getpid()                                = 408
getuid32()                              = 0
getpid()                                = 408
time(NULL)                              = 1137015810
getpid()                                = 408
stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=837, ...}) = 0
socketpair(PF_FILE, SOCK_STREAM, 0, [3, 4]) = 0
fork()                                  = 409
--- SIGCHLD (Child exited) @ 0 (0) ---
rt_sigaction(SIGTERM, {0x8049900, [TERM], SA_RESTORER|SA_RESTART,
0x256428}, {SIG_DFL}, 8) = 0
rt_sigaction(SIGINT, {0x8049900, [INT], SA_RESTORER|SA_RESTART,
0x256428}, {SIG_DFL}, 8) = 0
rt_sigaction(SIGCHLD, {0x8049900, [CHLD], SA_RESTORER|SA_RESTART,
0x256428}, {SIG_DFL}, 8) = 0
rt_sigaction(SIGHUP, {0x8049900, [HUP], SA_RESTORER|SA_RESTART,
0x256428}, {SIG_DFL}, 8) = 0
close(4)                                = 0
getpid()                                = 408
poll([{fd=3, events=POLLIN, revents=POLLIN|POLLHUP}], 1, -1) = 1
read(3, "", 65535)                      = 0
write(2, "dispatch_imsg in main: pipe clos"..., 35dispatch_imsg in main:
pipe closed
) = 35
rt_sigaction(SIGCHLD, {SIG_DFL}, {0x8049900, [CHLD],
SA_RESTORER|SA_RESTART, 0x256428}, 8) = 0
kill(409, SIGTERM)                      = 0
wait4(-1, NULL, 0, NULL)                = 409
wait4(-1, NULL, 0, NULL)                = -1 ECHILD (No child processes)
write(2, "Terminating\n", 12Terminating
)           = 12
exit_group(0)                           = ?
Process 408 detached


If you are interrested in more info (full strace, kernel config) please
mail me and i will send it directly to you.

Greetings
Bernhard


Am Mi 11.01.2006 15:19 schrieb Solar Designer <solar@...nwall.com>:

> On Wed, Jan 11, 2006 at 07:14:56AM +0100, Bernhard Fischer wrote:
> > Someone mailed problems with openntp last Dez.
> 
> Those were not "real" problems. Some log file messages were confusing,
> that's all.
> 
> > Now i am in the same
> > situation. Right after starting the daemon the forked child
> > terminates
> > and that's it. Strace does not show me any usefull hints
> > (socketpair()
> > is suspected for me to fail in this situation?).
> > 
> > The very strange thing is: i do have 2 machines (mostly the same
> > hardware and package releases) and only on one of them i can startup
> > ntpd.
> 
> Now this does sound like a real problem.  Please place your strace
> output on a private URL and post that.  We'll have a look.
> 
> Thanks,
> 
> -- 
> Alexander Peslyak <solar at openwall.com>
> GPG key ID: B35D3598 fp: 6429 0D7E F130 C13E C929 6447 73C3 A290 B35D
> 3598
> http://www.openwall.com - bringing security into open computing
> environments

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.