samba - Mar 2005 - Problem with 3.0.10 and 3.0.11 with 1 smbd process using99% cpu

I have this same problem on 3.0.10, and I also "fixed" it by deleting
the tdbs. My problem, had nothing to do with printing, it was happening
once I started winbindd with security = ADS. 

Jerry, what would we need to do to track this bug down in the tdb code?
I have logs at level 10, ltrace output, and a backtrace.

Do you think that there were changes made to the tdb code in 3.0.12 that
might fix this?

		-Marc
> -----Original Message-----
> From: samba-bounces+marc_kaplan=adaptec.com@lists.samba.org
[mailto:samba-
> bounces+marc_kaplan=adaptec.com@lists.samba.org] On Behalf Of John C.
> Hennessy
> Sent: Tuesday, March 01, 2005 12:07 PM
> To: samba@lists.samba.org
> Cc: john@mckissock.com
> Subject: Re: [Samba] Problem with 3.0.10 and 3.0.11 with 1 smbd
process
> using99% cpu
> 
> Just to let everyone know I was able to fix the problem. I cleaned out
> the tdb files which appearntly had been corrupted.
> 
> -John
> 
> John C. Hennessy wrote:
> 
> > I've been having problems since updating to samba 3.0.10 on Debian
3.1
> > Below is the output of ltrace and gdb on the offending smbd process.
> > I tried upgrading to 3.0.11 and the problem still exists. Anyone
have
> > any suggestions?
> >
> >
> > [ltrace output]
> > After about 20 seconds on the processes ltrace loops this
> >
> > iconv(0x82cecc8, 0xbffff218, 0xbffff21c, 0xbffff220, 0xbffff224) -1
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0, 0, 0, 0)                     = 0
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0xbffff218, 0xbffff21c, 0xbffff220, 0xbffff224) -1
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0, 0, 0, 0)                     = 0
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0xbffff218, 0xbffff21c, 0xbffff220, 0xbffff224) -1
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0, 0, 0, 0)                     = 0
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0xbffff218, 0xbffff21c, 0xbffff220, 0xbffff224) -1
> > __errno_location()                               = 0x403ac560
> > iconv(0x82cecc8, 0, 0, 0, 0)                     = 0
> > __errno_location()                               = 0x403ac560
> >
> > [gdb backtrace]
> > Attaching to program: /usr/sbin/smbd, process 10657
> > <snip>
> > 0x40202cf9 in memcpy () from /lib/tls/libc.so.6
> > (gdb) bt
> > #0  0x40202cf9 in memcpy () from /lib/tls/libc.so.6
> > #1  0x081ac059 in tdb_set_lock_alarm ()
> > #2  0x081ac20d in tdb_set_lock_alarm ()
> > #3  0x081ad49b in tdb_exists ()
> > #4  0x081ad6e3 in tdb_traverse ()
> > #5  0x081b4a79 in pjob_delete ()
> > #6  0x081b4f61 in pjob_delete ()
> > #7  0x081a450b in message_dispatch ()
> > #8  0x081b5186 in start_background_queue ()
> > #9  0x081ffd62 in main ()
> > (gdb)
> >
> > John C. Hennessy
> > President/CTO
> > HNK Technology Solutions, Inc.
> >
> >
> >
> 
> 
> --
> To unsubscribe from this list go to the following URL and read the
> instructions:  https://lists.samba.org/mailman/listinfo/samba

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Kaplan, Marc wrote:
| I have this same problem on 3.0.10, and I also "fixed"
| it by deleting the tdbs. My problem, had nothing to do
| with printing, it was happening once I started winbindd
| with security = ADS.
|
| Jerry, what would we need to do to track this bug down
| in the tdb code? I have logs at level 10, ltrace output,
| and a backtrace.
|
| Do you think that there were changes made to the tdb
| code in 3.0.12 that might fix this?

Jeremy recently applied a patch to break from an inifinite
loop when tranversing a tdb.  This sounds like it might apply
here.  That was post 3.0.12pre1 IIRC.  Try the latest SAMBA_3_0
code.




cheers, jerry
====================================================================Alleviating
the pain of Windows(tm)      ------- http://www.samba.org
GnuPG Key                ----- http://www.plainjoe.org/gpg_public.asc
"I never saved anything for the swim back."     Ethan Hawk in Gattaca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCLHKQIR7qMdg1EfYRAsfAAKCcOzgAchxKiFonVhUKJvCUSzQr5gCfTpLo
UiUGgxMYHJHHNi1MeDcSB74=3kJv
-----END PGP SIGNATURE-----

On Mon, Mar 07, 2005 at 09:26:08AM -0600, Gerald (Jerry) Carter
wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Kaplan, Marc wrote:
> | I have this same problem on 3.0.10, and I also "fixed"
> | it by deleting the tdbs. My problem, had nothing to do
> | with printing, it was happening once I started winbindd
> | with security = ADS.
> |
> | Jerry, what would we need to do to track this bug down
> | in the tdb code? I have logs at level 10, ltrace output,
> | and a backtrace.
> |
> | Do you think that there were changes made to the tdb
> | code in 3.0.12 that might fix this?
> 
> Jeremy recently applied a patch to break from an inifinite
> loop when tranversing a tdb.  This sounds like it might apply
> here.  That was post 3.0.12pre1 IIRC.  Try the latest SAMBA_3_0
> code.
But that only works for a corrupt tdb. If you have that you also
have other problems...

Jeremy.