Problem creating GPO samba4 beta6

[steve]

On 28/08/12 14:23, Andrew Bartlett wrote:
> On Tue, 2012-08-28 at 14:17 +0200, Dieter Modig wrote:
>> Hi again!
>>
>> ----- Ursprungligt meddelande -----
>>
>>> Från: "Andrew Bartlett" <abartlet at samba.org>
>>> Till: "Dieter Modig" <dieter.m at inputinterior.se>
>>> Kopia: samba-technical at lists.samba.org
>>> Skickat: måndag, 27 aug 2012 23:51:01
>>> Ämne: Re: Problem creating GPO samba4 beta6
>>
>>> On Mon, 2012-08-27 at 11:05 +0200, Dieter Modig wrote:
>>>> Hi!
>>>>
>>>> Since one of the updates on our Samba4 environment (from alpa17 to
>>>> beta4) we can't seem to create new GPOs using windows GPO manager.
>>>> We can edit the existing ones but not create new ones. When trying
>>>> to create a new GPO I get an error message saying something like
>>>> "File/object not found". Removing this message and trying again
>>>> (with the same name) gets me a different error message saying
>>>> "Access denied". No logs seem to catch this so there is no further
>>>> debug info.
>>>>
>>>> We could however create a GPO from the linux command line using
>>>> samba-tool and that works but this GPO can't be used from windows
>>>> GPO manager.
>>>>
>>>> It seems to be a problem with access rights. Looking in the folder
>>>> /usr/local/samba/var/locks/sysvol/input.se/Policies/ we can see
>>>> that existing policies have different owners and groups. Some of
>>>> them have local linux users as owners and some of them have users
>>>> from the domain as owners. The theory that the problem is all
>>>> rights based is corraborated by the fact that sometimes we get an
>>>> error message saying "The permissions for this GPO in the sysvol
>>>> folder are inconsistent with those in active directory" and the
>>>> option to repair this. It does not, however, help to repair it :(
>>>>
>>>> What are the rights supposed to be? Is the Policies folder supposed
>>>> to be owned by local linux user (which is running the processes)
>>>> or a domain user (which is the one accessing the files)? Are there
>>>> any checks/fixes that we can run in order to see if there are
>>>> errors in the setup? This was working just fine before updating to
>>>> the beta release so has there been any changes in how the rights
>>>> are suppose to be set?
>>
>>> G'day,
>>
>>> I've been working to make this handle much better, and the beta7 due
>>> today (and current master) will work much better for you.
>>
>>> In particular, the new tool 'samba-tool ntacl sysvolreset' will set
>>> posix permissions to match the NT ACL, the lack of which is I hope
>>> the
>>> cause of your problems.
>>
>>> Let me know if it does or doesn't help, and I'll see what I can do.
>>
>> Uumm... I was very excited about the beta7 so it was downloaded, compiled and installed. And then we can't the server online anymore at all! :( Log.samba was not very forthcoming other than:
>>
>> [2012/08/28 12:21:38, 0] ../lib/util/fault.c:73(fault_report)
>> INTERNAL ERROR: Signal 11 in pid 1498 (4.0.0beta7)
>> Please read the Trouble-Shooting section of the Samba HOWTO
>> [2012/08/28 12:21:38, 0] ../lib/util/fault.c:75(fault_report)
>> ===============================================================
>> [2012/08/28 12:21:38, 0] ../lib/util/fault.c:144(smb_panic_default)
>> PANIC: internal error
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1509 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1508 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1507 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1506 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1505 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1504 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1503 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1502 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1501 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1499 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1497 on SIGTERM
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:115(sig_term)
>> SIGTERM: killing children
>> [2012/08/28 12:29:00, 0] ../source4/smbd/server.c:120(sig_term)
>> Exiting pid 1045 on SIGTERM
>> [2012/08/28 13:16:37, 0] ../source4/smbd/server.c:369(binary_smbd_main)
>> samba version 4.0.0beta7 started.
>> Copyright Andrew Tridgell and the Samba Team 1992-2012
>> [2012/08/28 13:16:37, 0] ../source4/smbd/server.c:475(binary_smbd_main)
>> samba: using 'standard' process model
>> [2012/08/28 13:16:39, 0] ../lib/util/fault.c:72(fault_report)
>>
>> So we had to go back to beta6 and now it's once again online. Next step would be to install beta7 from scratch using a clean database in order to see if the problem is the beta release or our database. Any suggestions?
>
> If your database can make us segfault, then please get us the details of
> the fault (the backtrace).
>
> Andrew Bartlett
>
Hi Dieter
We have the same
INTERNAL ERROR: Signal 11
on both the 4.0.0beta7 tarball and the next git after that: 
4.0.0beta8-GIT-11a5646, and the next git after that too (just rolled 
back so can't remember it but about 3 hours ago, 14:00 CET)

Do you know how to do the backtrace that Andrew asks for?

I'm currently doing a totally new install of the beta7 tarball onto bare 
metal. Maybe it is the beta itself?

Cheers,
Steve