[Samba] Not enough storage space error

Fri Aug 4 16:44:39 UTC 2017

Hello Louis,

You are welcome.

As Andrew suggested, we are now monitoring the memory and written a 
script to dump the memory consumption information at given interval of 
20 minutes. We will share the results here.

One thing we observed today. None connected to RSAT today and we kept 
adding the PCs to Domain. All versions of Windows - XP with SP3, Windows 
7 Professional and Ultimate (32 and 64Bit) and Windows 10 Professional 
and Enterprise. Everything worked just fine. We were not cut off by 
Samba at any time. However, we have now given instructions to all not to 
touch the RSAT either on Windows 7 Professional or on Windows Server 
2008 R2 when we are adding PCs to domain. If the problem does not 
repeat, then we feel probably we are heading towards some conclusion.

In our case, there is all the possibility that I connect to RSAT on 
Windows Server 2008 R2 from one location and another personal at the 
same time connects to RSAT from Windows 7 Professional. While I keep 
observing the Computer Accounts and DNS entries, the other person works 
on OU reorganization etc. This could be leading to this sudden 
disconnection as DB may go inconsistent. I am not sure.

However, I will check the error from Windows Events and share the 
details here. As Marc suggested, earlier, I will also execute dbcheck 
cross-ncs and post the results here along with memory utilization data 
over a period of time.

Just to give all information about our setup:

OS : CentOS 7.3 (1611)

ADDC : 4.6.5

No. of DC : 2 (1 in Physical Server with 2 X Octacore CPU, 64 GB Memory, 
300 GB HDD & another Virtual Machine, 8 vCPU, 16 GB Memory, 50 GB HDD)

Total number of users accounts added : around 5000

Total Computers added to Domain : around 500 as of now.

Target is to achieve close to 8000 users and equal number of PCs.

-- 

Thanks & Regards,

Anantha Raghava

Do not print this e-mail unless required. Save Paper & trees.

On 04/08/17 3:55 PM, L.P.H. van Belle wrote:
> Hai,
>
> Im joining this one if you dont mind.
> My setup:
> Debian 8 (Jessie)
> ADDC's  : samba 4.6.6
> Member with profiles : samba 4.6.6
> This problem is here seen also on a Windows7 Pro 64bit client.
> The client here, is related to the user profile, this one is having a corrupted user profile in my case.
> Due to this, my GPO fails and i see event id: 4098 , error code : 0x8007000E insufficions space ....
>
> I also seen some hotfixes related to dotnet 3.0 for Win 7, but im not sure yet for Anantha what the problem is.
>
> Also dont see this on my other computers, checked.
> Windows 7 Pro 64bit ( other users )
> Windows 10 Pro 64bit. Works fine also.  ( builds 15xx upto 17xx )
>
>
> @Anantha, can you check if you see in the windows event logs with warnings and errors.
> Can you post these? Event id NR. and error codes that all i need.
>
> Imho, this is not samba that causing the problem but something in windows.
>
>
> Greetz,
>
> Louis
>
>
>
>> -----Oorspronkelijk bericht-----
>> Van: samba [mailto:samba-bounces at lists.samba.org] Namens
>> Anantha Raghava via samba
>> Verzonden: vrijdag 4 augustus 2017 6:03
>> Aan: Andrew Bartlett; samba at lists.samba.org
>> Onderwerp: Re: [Samba] Not enough storage space error
>>
>> Hello Andrew,
>>
>> Good Morning..
>>
>> Thanks for quick response.
>>
>> We will monitor the memory usage and intimate you accordingly.
>>
>> However, I doubt using the 4.7rc in this case. Also, creating
>> a test environment with 4.7rc, simulating the current process
>> (Adding new domain member PCs and users, reorganizing the
>> Groups and OU) is very difficult. Will definitely give it a try.
>>
>> -- 
>>
>> Thanks & Regards,
>>
>>
>> Anantha Raghava
>>
>>
>> Do not print this e-mail unless required. Save Paper & trees.
>>
>> On 04/08/17 8:44 AM, Andrew Bartlett wrote:
>>> On Fri, 2017-08-04 at 06:34 +0530, Anantha Raghava wrote:
>>>> Hello Andrew,
>>>>   From the day one, we set the log level to 3. Log size is
>> really huge.
>>>> However, I did not see some thing very interesting and related to
>>>> this error. It is all to do with user login, information,
>> etc. Should
>>>> I have to increase the log level to get more info?
>>>> One thing we noticed. When we restart the samba-ad-dc service, it
>>>> allows us to add new users, add new computers to domain, and even
>>>> move the user from one OU to another. However, from last
>> two days, at
>>>> around evening (around 4:45 PM), samba all of a sudden
>> cuts off all
>>>> users from domain and even fails to list the domain controllers in
>>>> RSAT. If we restart samba-ad-dc, it starts working properly.
>>>> Now we need to see, what is that event that is causing
>> this cut off
>>>> all of a sudden? Is it a shear coincidence? Or is it
>> something to do
>>>> with the large number of PCs are now getting added from different
>>>> places? Incidentally, we use the same domain admin account (limited
>>>> administrator) to add PCs to domain. Is this causing some problem?
>>> Can you track the memory use of samba over time?
>>>
>>> You say elsewhere that there is plenty of RAM free, but is
>> that still
>>> the case just before it blows up?
>>>
>>> Can you try Samba 4.7rc, where we made LDAP multi-process?
>> That would
>>> keep any memory leak more isolated, which might help us find it.
>>>
>>> You can also look for some clues in the output of:
>>>
>>>    smbcontrol $PID pool-usage
>>>
>>> It is very verbose, but careful analysis over time might
>> give a clue
>>> where the memory is.
>>>
>>> Thanks,
>>>
>>> Andrew Bartlett
>>>
>>>> --
>>>>
>>>> Thanks & Regards,
>>>>
>>>> Anantha Raghava
>>>>
>>>> Do not print this e-mail unless required. Save Paper & trees.
>>>> On 03/08/17 8:09 AM, Andrew Bartlett wrote:
>>>>> On Wed, 2017-08-02 at 21:09 +0530, Anantha Raghava via
>> samba wrote:
>>>>>> Hi,
>>>>>>
>>>>>> We are using Samba AD DC (Version 4.6.5) on CentOS 7.3.
>> We have two
>>>>>> servers with BIND9 DNS and all are working just fine.
>>>>>>
>>>>>> Today, when we attempted to move couple of users from one OU to
>>>>>> another, Windows RSAT reported /*"Windows cannot move object.
>>>>>> Ther*//*e *//*is not enough storage space*//*"*/.
>>>>> I wouldn't take the error string literally.  There are too many
>>>>> layers of mapping going on.  LDAP does not contain such
>> an error, so
>>>>> think of it as a bad translation.
>>>>>
>>>>>> Our servers have 300 GB Storage more that 270GB is free on the
>>>>>> storage. At the moment we have added about 1000 users
>> and about 450
>>>>>> computers into domain. Overall, we will add about 7500 users and
>>>>>> equal no of computers to domain. I also noticed that
>> RSAT lost the
>>>>>> connection to domain all of a sudden. However, when I
>> restarted the
>>>>>> samba-ad-dc service, it reconnected. I am able to add
>> new computers
>>>>>> and new users into domain. But moving the user from one OU to
>>>>>> another is throwing above error.
>>>>>>
>>>>>> I believe the DB is a 32 bit one and should support upto 4 GB.
>>>>>> When
>>>>>> I
>>>>>> checked in /usr/local/samba/private folder the size of
>> smb.ldb file
>>>>>> is about 4.1 MB and rest of the ldb and tdb files are
>> around 1 MB
>>>>>> in size.
>>>>>>
>>>>>> Now, how do I correctly check the current DB size? Is the above
>>>>>> error a strange behaviour? How do we fix the above error?
>>>>> Turn up the logging on the AD DC and see what the real error and
>>>>> error string is.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Andrew Bartlett
>>>>>
>>>>    
>> --
>> To unsubscribe from this list go to the following URL and read the
>> instructions:  https://lists.samba.org/mailman/options/samba
>>