Bind is hanging on CentOS 4.4
Jeff Lightner
jlightner at water.com
Fri May 29 18:01:52 UTC 2009
At least now it shows all the threads on a single processor.
-----Original Message-----
From: bind-users-bounces at lists.isc.org [mailto:bind-users-bounces at lists.isc.org] On Behalf Of Jesse Cabral
Sent: Friday, May 29, 2009 1:57 PM
To: 'Jeffrey Reasoner'
Cc: bind-users at lists.isc.org
Subject: RE: Bind is hanging on CentOS 4.4
Importance: High
Good suggestion.
I just tried that and re-ran the ./configure --disable-threads
Then I killed the named pid and started named:
ps -Leo user,pid,ppid,lstart,lwp,nlwp,psr,args |egrep "LWP|named"
USER PID PPID STARTED LWP NLWP PSR COMMAND
named 14671 1 Fri May 29 13:56:41 2009 14671 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 14671 1 Fri May 29 13:56:41 2009 14672 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 14671 1 Fri May 29 13:56:41 2009 14673 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 14671 1 Fri May 29 13:56:41 2009 14674 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 14671 1 Fri May 29 13:56:41 2009 14675 5 0 /usr/sbin/named
-u named -t /var/named/chroot
root 14678 3402 Fri May 29 13:56:43 2009 14678 1 1 egrep LWP|named
Still appears to be threaded ?
Jesse Cabral
Solutions Engineer
Micro Technology Solutions, Inc.
Phone:508.324.9475 Fax:508.324.4477
21 Father DeValles Blvd, Suite 101, Fall River, MA 02723
www.mtsolutions.net
-----Original Message-----
From: Jeffrey Reasoner [mailto:jeff.reasoner at mail.hccanet.org]
Sent: Friday, May 29, 2009 1:38 PM
To: 'jcabral at mtsolutions.net'
Subject: RE: Bind is hanging on CentOS 4.4
Maybe 'make distclean' first and then rerun ./configure?
-----Original Message-----
From: bind-users-bounces at lists.isc.org [mailto:bind-users-bounces at lists.isc.
org] On Behalf Of Jesse Cabral
Sent: Friday, May 29, 2009 12:58 PM
To: 'Jeff Lightner'
Cc: bind-users at lists.isc.org
Subject: RE: Bind is hanging on CentOS 4.4
Importance: High
I can't seem to get bind reconfigured to run with no threads ? Anyone have
any thoughts ?
When I rerun ./configure --disable-threads the configure file does not
update.
Regards,
Jesse Cabral
Solutions Engineer
Micro Technology Solutions, Inc.
Phone:508.324.9475 Fax:508.324.4477
21 Father DeValles Blvd, Suite 101, Fall River, MA 02723
www.mtsolutions.net
-----Original Message-----
From: Jeff Lightner [mailto:jlightner at water.com]
Sent: Friday, May 29, 2009 9:03 AM
To: jcabral at mtsolutions.net
Subject: RE: Bind is hanging on CentOS 4.4
Yes. That's the way I would read it. Starting with -n 1 means it should
all be on one processor even though it is running multiple threads.
Try this:
ps -Leo user,pid,ppid,lstart,lwp,nlwp,psr,args |egrep "LWP|named"
USER PID PPID STARTED LWP NLWP PSR COMMAND
root 2833 1 Wed Jan 14 10:51:40 2009 2833 1 0 syslogd -m 0 -a
/var/named/chroot/dev/log
named 12622 1 Thu May 14 09:51:36 2009 12622 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 12622 1 Thu May 14 09:51:36 2009 12623 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 12622 1 Thu May 14 09:51:36 2009 12624 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 12622 1 Thu May 14 09:51:36 2009 12625 5 0 /usr/sbin/named
-u named -t /var/named/chroot
named 12622 1 Thu May 14 09:51:36 2009 12626 5 1 /usr/sbin/named
-u named -t /var/named/chroot
root 19429 19079 Fri May 29 08:59:39 2009 19429 1 0 egrep LWP|named
The above shows that I'm running 5 threads (NWLP column) of named and that 4
of them are on processor 0 and the final 1 is on processor 1 (psr column).
All are PID 12622 but the LWP (thread ID) is different for each.
-----Original Message-----
From: Jesse Cabral [mailto:jcabral at mtsolutions.net]
Sent: Thursday, May 28, 2009 4:55 PM
To: Jeff Lightner
Subject: RE: Bind is hanging on CentOS 4.4
Importance: High
Here is some more info:
I decided to start named with the -n 1 then
I ran ps -eLfc | grep named
named 26750 1 26750 4 TS 18 16:42 ? 00:00:00
/usr/sbin/named -u named -n 1 -t /var/named/chroot
named 26750 1 26751 4 TS 24 16:42 ? 00:00:15
/usr/sbin/named -u named -n 1 -t /var/named/chroot
named 26750 1 26752 4 TS 24 16:42 ? 00:00:00
/usr/sbin/named -u named -n 1 -t /var/named/chroot
named 26750 1 26753 4 TS 24 16:42 ? 00:00:03
/usr/sbin/named -u named -n 1 -t /var/named/chroot
4TS should be 4 threads. If I run
ps -p 26750 -T
PID SPID TTY TIME CMD
26750 26750 ? 00:00:00 named
26750 26751 ? 00:00:20 named
26750 26752 ? 00:00:00 named
26750 26753 ? 00:00:04 named
It appears to be running 4 threads correct ?
Jesse Cabral
Solutions Engineer
Micro Technology Solutions, Inc.
Phone:508.324.9475 Fax:508.324.4477
21 Father DeValles Blvd, Suite 101, Fall River, MA 02723
www.mtsolutions.net
-----Original Message-----
From: Jeff Lightner [mailto:jlightner at water.com]
Sent: Thursday, May 28, 2009 4:26 PM
To: jcabral at mtsolutions.net
Cc: bind-users at lists.isc.org
Subject: RE: Bind is hanging on CentOS 4.4
The output suggests named is running with threads (I see similar output on
my RHEL BIND server).
Try running ps -efLm |grep named to get a better look at it. Some time ago
they changed the way threads are done in Linux (one of the 2.4.x kernels).
The columns LWP and NWLP from the ps -eLf should show thread id and number
of threads respectively per the man page but based on what I'm seeing on my
RHEL system the latter is showing 0 when in fact there are multiple threads.
-----Original Message-----
From: Jesse Cabral [mailto:jcabral at mtsolutions.net]
Sent: Thursday, May 28, 2009 3:55 PM
To: Jeff Lightner
Cc: bind-users at lists.isc.org
Subject: RE: Bind is hanging on CentOS 4.4
Importance: High
Here is the output from ps -eLf
ps -eLf | grep named
named 32231 1 32231 0 7 May26 ? 00:00:00 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32232 0 7 May26 ? 00:15:04 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32236 0 7 May26 ? 00:15:04 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32237 0 7 May26 ? 00:14:58 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32238 0 7 May26 ? 00:15:06 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32239 0 7 May26 ? 00:00:07 /usr/sbin/named
-u named -t /var/named/chroot
named 32231 1 32240 0 7 May26 ? 00:12:40 /usr/sbin/named
-u named -t /var/named/chroot
I rebuild named with threads disabled and I still hand a hang afterward.
Jesse Cabral
Solutions Engineer
Micro Technology Solutions, Inc.
Phone:508.324.9475 Fax:508.324.4477
21 Father DeValles Blvd, Suite 101, Fall River, MA 02723
www.mtsolutions.net
-----Original Message-----
From: Jeff Lightner [mailto:jlightner at water.com]
Sent: Thursday, May 28, 2009 1:27 PM
To: jcabral at mtsolutions.net
Subject: RE: Bind is hanging on CentOS 4.4
The implementation of ps has a lot to do with whether you see threads.
You might try:
To get info about threads:
ps -eLf
ps axms
-----Original Message-----
From: bind-users-bounces at lists.isc.org [mailto:bind-users-bounces at lists.isc.
org] On Behalf Of Jesse Cabral
Sent: Thursday, May 28, 2009 1:16 PM
To: 'JINMEI Tatuya / 神明達哉'
Cc: bind-users at lists.isc.org
Subject: RE: Bind is hanging on CentOS 4.4
Importance: High
I have decided to rebuild named with threads disabled. I will update you on
the results. I am hoping that will remedy the issue.
Thanks for your support.
Jesse Cabral
Solutions Engineer
Micro Technology Solutions, Inc.
Phone:508.324.9475 Fax:508.324.4477
21 Father DeValles Blvd, Suite 101, Fall River, MA 02723
www.mtsolutions.net
-----Original Message-----
From: JINMEI Tatuya / 神明達哉 [mailto:Jinmei_Tatuya at isc.org]
Sent: Tuesday, May 26, 2009 3:14 PM
To: jcabral at mtsolutions.net
Subject: Re: Bind is hanging on CentOS 4.4
At Tue, 26 May 2009 09:23:22 -0400,
"Jesse Cabral" <jcabral at mtsolutions.net> wrote:
> When I run a ps ax | grep named
>
> I only see a single thread for named ? If I was using multi-threads it
> should show an additional thread for each instance correct ?
It depends on the ps implementation. To be sure, you should start
named with the -g command line option and see initial log messages
on stderr. If you see something like this:
26-May-2009 12:11:39.619 found 1 CPU, using 1 worker thread
threads are enabled. If the threads are enabled, the best way to
eliminate the possibility of dead lock is to rebuild named
--disable-threads. If you cannot do it, maybe you want to try
invoking named with the '-n 1' command line option.
---
JINMEI, Tatuya
Internet Systems Consortium, Inc.
p.s. if you don't mind, please cc follow up messages to the bind-users
list. Then you can expect more help from others. I'm just back from
vacation with so many email backlogs and my responses may be overly
delayed.
_______________________________________________
bind-users mailing list
bind-users at lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
Please consider our environment before printing this e-mail or attachments.
----------------------------------
CONFIDENTIALITY NOTICE: This e-mail may contain privileged or confidential
information and is for the sole use of the intended recipient(s). If you are
not the intended recipient, any disclosure, copying, distribution, or use of
the contents of this information is prohibited and may be unlawful. If you
have received this electronic transmission in error, please reply
immediately to the sender that you have received the message in error, and
delete it. Thank you.
----------------------------------
_______________________________________________
bind-users mailing list
bind-users at lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
_______________________________________________
bind-users mailing list
bind-users at lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
More information about the bind-users
mailing list