Another RE: 2.2.3pre winbindd crash

Noel Kelly nkelly at tarsus.co.uk
Wed Jan 9 03:18:05 GMT 2002


Just had winbind crash again - same rid 512 as before.  How can I track down
the reference to this rid in the domain ?

Also these messages about being unable to get the domain sid are clearly not
healthy but wbinfo -t tells me that 'secret is good'.

My panic action failed to open as well (the X display is correct) but I'm
not sure this is going to work anyway as the winbind process has probably
disappeared by then:

panic action = /usr/bin/X11/xterm -display :1.0 -e gdb
/usr/local/samba/bin/winbindd


Log excerpt:

[2002/01/09 10:56:11, 0]
nsswitch/winbindd_group.c:winbindd_getgrnam_from_group(220)
  getgrname_from_group(): could not get domain sid for domain @UK
[2002/01/09 10:57:56, 1] nsswitch/winbindd_group.c:fill_grent_mem(87)
  fill_grent_mem(): could not lookup membership for group rid 512 in domain
UK
[2002/01/09 10:57:56, 0] lib/fault.c:fault_report(38)
  ===============================================================
[2002/01/09 10:57:56, 0] lib/fault.c:fault_report(39)
  INTERNAL ERROR: Signal 11 in pid 2748 (2.2.3-pre)
  Please read the file BUGS.txt in the distribution
[2002/01/09 10:57:56, 0] lib/fault.c:fault_report(41)
  ===============================================================
[2002/01/09 10:57:56, 0] lib/util.c:smb_panic(1068)
  PANIC: internal error


-----Original Message-----
From: jra at samba.org [mailto:jra at samba.org]
Sent: 08 January 2002 20:35
To: Noel Kelly
Cc: 'samba at lists.samba.org'
Subject: Re: 2.2.3pre winbindd crash


On Tue, Jan 08, 2002 at 05:15:24PM -0000, Noel Kelly wrote:
> The second thing was a little more serious in that winbind blew-up just
> after lunch. Before I get into that I should point out that I am running
> winbind with the same winbindd_idmap.tdb from 2.2.2 to obvioulsy keep my
> user mappings but I am not sure if, or how, this affects things with
2.2.3.
> I have attached the relevant part of the winbind log below.  The actual
> event saw the winbind process consume 106% (!?) of the memory for about
five
> minutes before finally disappearing.

Can you add a panic action to smb.conf so we can catch the
backtrace in winbindd if this happens again please ?

Thanks,

	Jeremy.




More information about the samba mailing list