syslogd causes hangs
Terry R Linhardt
linhardt at swbell.net
Wed Oct 8 03:44:42 UTC 2003
Okay...I can be dumb about this. :)
If I do: strace /sbin/syslogd -m 0 ...
the strace exits after a second or so, although syslogd continues to
run. Following is what strace captures before it exits.
Does this help?:
===========Following is output from strace /sbin/syslogd -m 0======
execve("/sbin/syslogd", ["/sbin/syslogd"], [/* 21 vars */]) = 0
2 uname({sys="Linux", node="chastain", ...}) = 0
3 brk(0) = 0x9b4e000
4 open("/etc/ld.so.preload", O_RDONLY) = -1 ENOENT (No such
file or dir ectory)
5 open("/etc/ld.so.cache", O_RDONLY) = 3
6 fstat64(3, {st_mode=S_IFREG|0644, st_size=67825, ...}) = 0
7 old_mmap(NULL, 67825, PROT_READ, MAP_PRIVATE, 3, 0) = 0xbf58d000
8 close(3) = 0
9 open("/lib/tls/libc.so.6", O_RDONLY) = 3
10 read(3,
"\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0 at X\1\000"..., 51
2) = 512
11 fstat64(3, {st_mode=S_IFREG|0755, st_size=1571976, ...}) = 0
12 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1 , 0) = 0xbf58c000
13 old_mmap(NULL, 1275244, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0)
= 0x1110 00
14 old_mmap(0x243000, 12288, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED, 3 , 0x132000) = 0x243000
15 old_mmap(0x246000, 9580, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP _ANONYMOUS, -1, 0) = 0x246000
16 close(3) = 0
17 set_thread_area({entry_number:-1 -> 6, base_addr:0xbf58c760,
limit:10485 75, seg_32bit:1, contents:0, read_exec_only:0,
limit_in_pages:1, seg_not _present:0, useable:1}) = 0
18 munmap(0xbf58d000, 67825) = 0
19 getpid() = 2921
20 chdir("/") = 0
21 brk(0) = 0x9b4e000
22 brk(0x9b6f000) = 0x9b6f000
23 brk(0) = 0x9b6f000
24 open("/var/run/syslogd.pid", O_RDONLY) = 3
25 fstat64(3, {st_mode=S_IFREG|0600, st_size=5, ...}) = 0
26 mmap2(NULL, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0 ) = 0xbf59d000
27 read(3, "2917\n", 4096) = 5
28 close(3) = 0
29 munmap(0xbf59d000, 4096) = 0
30 getpid() = 2921
31 kill(2917, SIG_0) = -1 ESRCH (No such
process)
32 clone(child_stack=0,
flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCH LD,
child_tidptr=0xbf58c7a8) = 2922
33 rt_sigaction(SIGTERM, {0x804bb30, [TERM],
SA_RESTORER|SA_RESTART, 0x138a 18}, {SIG_DFL}, 8) = 0
34 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
35 rt_sigaction(SIGCHLD, NULL, {SIG_DFL}, 8) = 0
36 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
37 nanosleep({300, 0}, 0xbff0c8c4) = -1 EINTR (Interrupted
system c all)
38 --- SIGTERM (Terminated) @ 0 (0) ---
39 exit_group(0) =
==========================End Output =============================
By the way, I have another machine on which I CANNOT duplicate this
problem. But, on my "problem" machine syslogd is definitely the problem.
It seems to somehow interfere with Nautilus. As noted before, mail and
browser response are okay.
Also, system load is not an issue. Machine remains at ~85% idle when I
am having "issues".
Terry
On Mon, 2003-10-06 at 23:15, Bill Nottingham wrote:
> Terry R Linhardt (linhardt at swbell.net) said:
> > Tonight I installed updates to Beta 2. The Fedora splash screen simply
> > hung. (Never went to a desktop) I got to a text screen, and did some
> > poking around. Finally discovered that if I *killed* the syslogd process
> > my splash screen immediately "processed" and I got a graphical desktop.
> > Also, graphical access to my home directory worked like a charm. But of
> > course my syslogd daemon is now dead. Oh yes, if I manually restart the
> > daemon, such things as the graphical access to my home directory once
> > again go into the dumper. (Although command line manipulation of the
> > directory structure works without a problem).
> >
> > Feedback would be appreciated.
>
> What's sysklogd doing if you strace it?
>
> Bill
>
>
> --
> fedora-test-list mailing list
> fedora-test-list at redhat.com
> http://www.redhat.com/mailman/listinfo/fedora-test-list
--
Terry R Linhardt <linhardt at swbell.net>
More information about the fedora-test-list
mailing list