(2005.10.21, 13:39)
Ok, if I understand you you're re-exporting an NFS mount from the
Linux box to the Windows clients via Samba. Looks to me like the
kernel share mode code isn't working over NFS properly. Can you try
removing the define #define HAVE_KERNEL_SHARE_MODES 1 from
On Fri, Oct 21, 2005 at 01:40:16PM -0400, Benoit Gauthier wrote:
(2005.10.21, 13:39)
Ok, if I understand you you're re-exporting an NFS mount from the
Linux box to the Windows clients via Samba. Looks to me like the
kernel share mode code isn't working over NFS properly. Can you try
(2005.10.21, 14:34)
Ok, if I understand you you're re-exporting an NFS mount from the
Linux box to the Windows clients via Samba. Looks to me like the
kernel share mode code isn't working over NFS properly. Can you try
removing the define #define HAVE_KERNEL_SHARE_MODES 1 from
Jerry, I'm sorry I still could not provide you with the traces you asked
for, but in the meantime we have not had any fatal crash.
The one thing that changed on our network since the weekend is that we
shutdown an old DNS server on a different machine which held partially
incorrect data (but
(2005.10.19, 07:36)
OK. I've figured out how to do it. Which of the numerous Samba
processes should I do this on though?
On one of those you've tried to kill with a normal kill command but
hasn't died
Here is such an output.
Benoît
Attaching to program: /usr/sbin/smbd, process 3456
On Wed, Oct 19, 2005 at 07:37:05AM -0400, Benoit Gauthier wrote:
(2005.10.19, 07:36)
OK. I've figured out how to do it. Which of the numerous Samba
processes should I do this on though?
On one of those you've tried to kill with a normal kill command but
hasn't died
Here is such
(2005.10.19, 13:10)
The flock is called only if HAVE_KERNEL_SHARE_MODES is set to
true. What kernel are you running on ?
Does 2.6.11 make sense?
Do you know if any NFS client have files open on the Samba share
also ?
The configuration is like this:
4 external computers run NFS servers.
On Wed, Oct 19, 2005 at 01:16:57PM -0400, Benoit Gauthier wrote:
(2005.10.19, 13:10)
The flock is called only if HAVE_KERNEL_SHARE_MODES is set to
true. What kernel are you running on ?
Does 2.6.11 make sense?
Do you know if any NFS client have files open on the Samba share
also ?
(2005.10.19, 16:59)
Ok, if I understand you you're re-exporting an NFS mount from the
Linux box to the Windows clients via Samba. Looks to me like the
kernel share mode code isn't working over NFS properly. Can you try
removing the define #define HAVE_KERNEL_SHARE_MODES 1 from
On Wed, Oct 19, 2005 at 05:00:51PM -0400, Benoit Gauthier wrote:
(2005.10.19, 16:59)
Ok, if I understand you you're re-exporting an NFS mount from the
Linux box to the Windows clients via Samba. Looks to me like the
kernel share mode code isn't working over NFS properly. Can you try
(2005.10.18, 10:48)
Never kill smbd with -9. It can leave important tdb files in a
corrupt state.
So, what is the right way to kill smbd altogether?
Until a solution is found to the current problem, I need to start it
afresh so that my back-ups will take place.
Thanks in advance.
Benoît
On Tue, Oct 18, 2005 at 10:49:41AM -0400, Benoit Gauthier wrote:
(2005.10.18, 10:48)
Never kill smbd with -9. It can leave important tdb files in a
corrupt state.
So, what is the right way to kill smbd altogether?
Just use killall smbd. If it doesn't terminate that's a bug and
we need
(2005.10.18, 12:30)
So, what is the right way to kill smbd altogether?
Just use killall smbd. If it doesn't terminate that's a bug and
we need to investigate why it's hanging/looping.
Well, it doesn't do any good here. All processes are still active.
What next?
Benoît
--
To unsubscribe
On Tue, Oct 18, 2005 at 12:31:23PM -0400, Benoit Gauthier wrote:
(2005.10.18, 12:30)
So, what is the right way to kill smbd altogether?
Just use killall smbd. If it doesn't terminate that's a bug and
we need to investigate why it's hanging/looping.
Well, it doesn't do any good here.
(2005.10.17, 08:11)
Will one of the long lived processes respond to kill -TERM ?
All processes die after a killall -9 smbd even though the oldest one
takes several seconds to do so.
I have additional problems which appear to be associated with version
3.0.20b since I did not experience them
On Mon, Oct 17, 2005 at 08:17:43AM -0400, Benoit Gauthier wrote:
(2005.10.17, 08:11)
Will one of the long lived processes respond to kill -TERM ?
All processes die after a killall -9 smbd even though the oldest one
takes several seconds to do so.
Never kill smbd with -9. It can leave
creating multiple isntances of the deamon
On Mon, Oct 17, 2005 at 08:17:43AM -0400, Benoit Gauthier wrote:
(2005.10.17, 08:11)
Will one of the long lived processes respond to kill -TERM ?
All processes die after a killall -9 smbd even though the
oldest one
takes several seconds to do so
- Original Message -
From: Jeremy Allison [EMAIL PROTECTED]
To: Benoit Gauthier [EMAIL PROTECTED]; [EMAIL PROTECTED]
Cc: samba@lists.samba.org; Gerald (Jerry) Carter [EMAIL PROTECTED]
Sent: Monday, October 17, 2005 11:29 AM
Subject: Re: [Samba] Samba creating multiple isntances
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jeremy Allison wrote:
On Mon, Oct 17, 2005 at 08:17:43AM -0400, Benoit Gauthier wrote:
(2005.10.17, 08:11)
Will one of the long lived processes respond to kill -TERM ?
All processes die after a killall -9 smbd even though the oldest one
takes
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Barry, Christopher wrote:
is simply 'killall smbd' the preferre invocation?
Yes. But specifically the TERM signal allows for a
clean shutdown if possible.
cheers, jerry
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.0 (GNU/Linux)
Comment:
On Mon, Oct 17, 2005 at 10:44:56AM -0500, Gerald (Jerry) Carter wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jeremy Allison wrote:
On Mon, Oct 17, 2005 at 08:17:43AM -0400, Benoit Gauthier wrote:
(2005.10.17, 08:11)
Will one of the long lived processes respond to kill -TERM ?
Will someone please get me enough information to be able to
do something about this?
Jerry, I will provide you with the information as soon as we have the next
crash. (I have not yet found a way to intentionally make samba crash.)
Meanwhile, if you lookup the old thread [Samba] smbd hung
Hello Benoit Gerald,
I have the same problem here on a Debian Sarge server. (I also found an
older posting from 2004 by the subject of [Samba] smbd hung processes -
Samba 3.0.7, which described this problem, but back then the thread ended
without a solution.)
Benoit, perhaps we can compare
(2005.10.14, 05:56)
I have the same problem here on a Debian Sarge server.
OK. I have a Fedora 4 installation.
- Do you use LDAP authentication?
No.
- Do you use the name serving cache nscd?
No.
- Do you use printing over Samba/CUPS?
No.
- How is your login script called? I use a vbs
: Re: AW: [Samba] Samba creating multiple isntances
of the deamon
(2005.10.14, 05:56)
I have the same problem here on a Debian Sarge server.
OK. I have a Fedora 4 installation.
- Do you use LDAP authentication?
No.
- Do you use the name serving cache nscd?
No.
- Do you use printing over
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Benoit Gauthier wrote:
| That client is a Windows 2000 desktop which is 99%
| reserved for backup purposes. It connects to a local
| Linux computer (192.168.0.99) via Samba (using a mapped
| drive, of course) and then addresses four NFS
| mounts to
(2005.10.14, 09:59)
Is the machine accessing the Samba server via unc path name? You
might want to try it as a mapped drive.
It is already accessed via a mapped drive.
Benoît
--
To unsubscribe from this list go to the following URL and read the
instructions:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Benoit Gauthier wrote:
| (2005.10.14, 09:59)
|
| Is the machine accessing the Samba server via unc path name? You
| might want to try it as a mapped drive.
|
| It is already accessed via a mapped drive.
Can you send me an ethereal trace of a backup
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Louis van Belle wrote:
| Hi, i think i have the same problem here also ,
|
| i had a problem 2 weeks ago, my samba panic-ed, but
| i lost my logs with the panic code in it.
..
| and i'm sure everybody logs out every day, and there are
| only 2
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Benoit Gauthier wrote:
| (2005.10.10, 14:25)
|
| I am running Samba 3.0.14a-2 on Fedora 4. I have
| no problem seeing the Linux file system from
| any of the Windows (2000, SP4) computers attached
| to the network.
|
| The problem is that smbd
(2005.10.13, 15:50)
Here is the netstat output for SMB. What does it tell you?
tcp0 0 0.0.0.0:139 0.0.0.0:*
LISTEN 0 71509858709/smbd
tcp0 0 0.0.0.0:445 0.0.0.0:*
LISTEN 0
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Benoit Gauthier wrote:
| (2005.10.13, 15:50)
|
| Here is the netstat output for SMB. What does it tell you?
CLOSE_WAIT means that the process is waiting for a socket to
shutdown. I think you have a client connecting and
disconnecting rapidly.
(2005.10.13, 20:32)
CLOSE_WAIT means that the process is waiting for a socket to
shutdown. I think you have a client connecting and disconnecting
rapidly. Network browsing perhaps?
What is the client at 192.168.0.199?
That client is a Windows 2000 desktop which is 99% reserved for backup
(2005.10.10, 14:25)
I am running Samba 3.0.14a-2 on Fedora 4. I have no problem seeing the
Linux file system from any of the Windows (2000, SP4) computers
attached to the network.
The problem is that smbd processes are created for no apparent reason
and never die. I noticed this when the
34 matches
Mail list logo
