runaway processes in 3.0a20
Bradley W. Langhorst
brad at langhorst.com
Wed Dec 11 14:36:01 GMT 2002
On Tue, 2002-12-10 at 13:31, Esh, Andrew wrote:
> Maybe post a couple of iterations of strace output? That would give us
> a clue what is being looped on.
i just tried to get an strace on a runaway samba...
here is the output of smbstatus with some information xxx'ed
notice pid 17205 and 19911
Samba version 2.999+3.0.alpha20-4 for Debian
PID Username Group Machine
--
14509 gregoire labusers mgregoire (2-cis331xxx.com)
14879 lauelab labusers schachman (schachmanxxx.com)
14434 lauelab labusers yphantis (yphantisxxx.com)
17962 tlaue labusers camdir (2-cis605xxx.com)
14547 slucius labusers sue_lucius (2-cis693xxx.com)
14509 gregoire labusers mgregoire (2-cis331xxx.com)
19911 sfchase labusers sue_chase (2-cis809xxx.com)
18049 lauelab labusers pickels (2-cis750xxx.com)
14509 gregoire labusers mgregoire (2-cis331xxx.com)
14411 bwlang labusers xxx.xxx.x5.113 (2-cis113xxx.com)
14410 khartman labusers xxx.xxx.x7.138 (2-cis738xxx.com)
17651 lauelab labusers pickels (2-cis750xxx.com)
17206 sfchase labusers sue_chase (2-cis809xxx.com)
14903 jsk labusers jon kingsbury (2-cis114xxx.com)
14410 khartman labusers xxx.xxx.x7.138 (2-cis738xxx.com)
19910 lauelab labusers oldmcae (2-cis63xxx.com)
19895 lauelab labusers newmcae (2-cis743xxx.com)
Service pid machine Connected at
--
IPC$ 14509 mgregoire Mon Dec 9 14:01:19 2002
IPC$ 17962 camdir Tue Dec 10 15:12:11 2002
data 19910 oldmcae Wed Dec 11 08:55:08 2002
lauelab 19911 sue_chase Wed Dec 11 08:57:18 2002
data 17206 sue_chase Tue Dec 10 11:30:41 2002
IPC$ 14509 mgregoire Wed Dec 11 09:10:27 2002
sfchase 19911 sue_chase Wed Dec 11 08:57:19 2002
lauelab 17206 sue_chase Tue Dec 10 11:21:47 2002
IPC$ 14547 sue_lucius Mon Dec 9 13:35:49 2002
gregoire 14509 mgregoire Wed Dec 11 09:12:49 2002
lab_phaser 18049 pickels Tue Dec 10 15:55:26 2002
IPC$ 14903 jon kingsbury Mon Dec 9 15:53:29 2002
data 14879 schachman Mon Dec 9 15:47:43 2002
netlogon 19911 sue_chase Wed Dec 11 08:57:16 2002
IPC$ 17206 sue_chase Tue Dec 10 10:39:51 2002
bwlang 14411 xxx.xxx.x5.113 Mon Dec 9 12:55:15 2002
IPC$ 14410 xxx.xxx.x7.138 Tue Dec 10 10:00:38 2002
netlogon 14509 mgregoire Wed Dec 11 09:12:47 2002
data 19895 newmcae Wed Dec 11 08:49:52 2002
profiles 14509 mgregoire Wed Dec 11 09:12:04 2002
IPC$ 19911 sue_chase Wed Dec 11 08:57:20 2002
data 17651 pickels Tue Dec 10 13:12:41 2002
sfchase 17206 sue_chase Tue Dec 10 10:39:51 2002
data 19911 sue_chase Wed Dec 11 08:57:55 2002
lauelab 14879 schachman Mon Dec 9 15:47:43 2002
IPC$ 14509 mgregoire Wed Dec 11 09:13:16 2002
slucius 14547 sue_lucius Mon Dec 9 13:34:30 2002
IPC$ 14410 xxx.xxx.x7.138 Mon Dec 9 12:55:28 2002
khartman 14410 xxx.xxx.x7.138 Tue Dec 10 10:00:07 2002
data 14434 yphantis Mon Dec 9 12:59:59 2002
Locked files:
Pid DenyMode Access R/W Oplock Name
--
19910 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/23HR2E.003 Wed Dec 11
09:06:02 2002
19910 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/23HR2E.003 Wed Dec 11
09:06:02 2002
19895 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/23HR1E.003 Wed Dec 11 09:03:06 2002
19895 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/23HR1E.003 Wed Dec 11 09:03:06 2002
14434 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
/data/XLIDATA/HIGGS.INT Wed Dec 11 09:09:20 2002
14434 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
/data/XLIDATA/HIGGS.EQU Wed Dec 11 09:09:20 2002
19910 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/2B.001 Wed Dec 11
09:05:56 2002
19910 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/2B.001 Wed Dec 11
09:05:56 2002
14547 DENY_NONE 0x2019f RDWR NONE
path_replaced2002gifts.xls Wed Dec 11 08:05:01 2002
14434 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replacedHIGGS.VEL Wed Dec 11 09:09:20 2002
14434 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced/HIGGS.SCN Wed Dec 11 09:09:20 2002
19895 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/1B.001 Wed Dec 11 09:04:33 2002
19895 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/1B.001 Wed Dec 11 09:03:02 2002
19910 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/2N.001 Wed Dec 11
09:05:59 2002
19910 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/BlackboxPrototype/121102/2N.001 Wed Dec 11
09:05:59 2002
19895 DENY_DOS 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/1N.001 Wed Dec 11 09:03:04 2002
19895 DENY_NONE 0x1 RDONLY EXCLUSIVE+BATCH
path_replaced10-100-7.5/121102/1N.001 Wed Dec 11 09:03:04 2002
here is the output of the first few lines of top
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
17206 sfchase 19 0 2932 2932 1964 R 97.8 0.7 39:02 smbd
17753 root 9 0 900 900 660 S 0.9 0.2 8:56 top
19848 root 13 0 904 904 668 R 0.9 0.2 0:00 top
14394 root 9 0 1548 1548 800 S 0.1 0.4 0:25 nmbd
1 root 8 0 480 480 416 S 0.0 0.1 0:14 init
so it looks like 17206 is a runaway...
hostname:~# strace -p 17206
hostname:~# kill -HUP 17206;strace -p 17206
--- SIGSTOP (Stopped (signal)) ---
hostname:~# kill 17206;strace -p 17206
--- SIGSTOP (Stopped (signal)) ---
hostname:~# kill -9 17206;strace -p 17206
+++ killed by SIGKILL +++
not very interesting output from strace...
am I doing something wrong?
thanks
brad
> --Original Message--
> From: Bradley W. Langhorst [mailto:brad at langhorst.com]
> Sent: Tuesday, December 10, 2002 10:47 AM
> To: samba-technical at samba.org
> Subject: runaway processes in 3.0a20
>
>
> I just installed the debian a20 packages
> (thanks steve and eloy)
>
> i'm posting to technical because it's about an alpha releas
>
>
> Once or twice a day i have to kill runaway processes (-9 required)
> they seem to as much cpu as is available
>
> I just saw one client with three connections to the server
> 1 was acting normally and the other two were runaway.
>
> before I go searching too deeply I'd like to eliminate a local
> misconfiguration or packaging problem...
> has anyone else seen this?
>
> i'll try to reproduce in a21 or 3.0's head if nobody has an idea about
> this...
>
> brad
>
>
> --
> Bradley W. Langhorst <brad at langhorst.com>
--
Bradley W. Langhorst <brad at langhorst.com>
More information about the samba-technical
mailing list