[Samba] [samba4] crash of winbind after "ls -l /usr/local/samba/var/locks/sysvol"

[François Lafont]

Hi,

Le 26/04/2013 11:38, Andrew Bartlett a écrit :

> Can you try that again with:
> 
> valgrind samba -i -M single > out 2>&1
> 
> I'm not worried about memory leaks (they don't cause crashes), and
> missing the '-' in front of -M probably means you didn't get the crash
> in the log you gave.  

Oops, sorry.

It's very strange, with your command, there is no problem in the log. I have tried this:

~# valgrind samba -i -M single > out 2>&1

During, the execution of samba, I have done this in another shell:

-----------------------------------------
~# wbinfo -u
Administrator
Guest
krbtgt
test1

~# ls -l /usr/local/samba/var/locks/sysvol/ # 1 minute'wait 
total 8
drwxrws---+ 4 root 3000000 4096 Apr 21 13:00 chezmoi.priv

~# wbinfo -u
Error looking up domain users
-----------------------------------------

And here my "out" file:

-----------------------------------------
==2377== Memcheck, a memory error detector
==2377== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al.
==2377== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info
==2377== Command: samba -i -M single
==2377== 
samba version 4.0.5 started.
Copyright Andrew Tridgell and the Samba Team 1992-2012
samba: using 'single' process model
==2377== 
==2377== HEAP SUMMARY:
==2377==     in use at exit: 4,601,429 bytes in 36,998 blocks
==2377==   total heap usage: 696,172 allocs, 659,174 frees, 113,032,156 bytes allocated
==2377== 
==2377== LEAK SUMMARY:
==2377==    definitely lost: 0 bytes in 0 blocks
==2377==    indirectly lost: 0 bytes in 0 blocks
==2377==      possibly lost: 4,283,837 bytes in 35,214 blocks
==2377==    still reachable: 317,592 bytes in 1,784 blocks
==2377==         suppressed: 0 bytes in 0 blocks
==2377== Rerun with --leak-check=full to see details of leaked memory
==2377== 
==2377== For counts of detected and suppressed errors, rerun with: -v
==2377== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 285 from 8)
-----------------------------------------

No problem ! It's very curious. I hope that we would more information with the thomas'log...

-- 
François Lafont