[Samba] sudden replication failures

[Antun Horvat]

Do a "strace -p pid_of_high_cpu_process > log.txt"  and report it here. 
I saw a bug in samba process which also eat cpu. The product of strace 
was that there was some internal fcntl lock/unlock loop on domain tdb file.

On 07/01/2014 09:13 PM, mourik jan heupink - merit wrote:
> Hi all,
>
> Samba (sernet) 4.1.7, and we're getting high cpu usage on dc1, and on 
> dc2 (my 'secondary' dc):
>
> root at dc2:~# samba-tool drs showrepl
> Default-First-Site-Name\DC2
> DSA Options: 0x00000001
> DSA object GUID: 5e93a102-2963-496a-af16-0c51eebb2e31
> DSA invocationId: 82b24c02-16cb-4ed1-b68b-a047c26886a9
>
> ==== INBOUND NEIGHBORS ====
>
> DC=DomainDnsZones,DC=samba,DC=domain,DC=com
>     Default-First-Site-Name\DC1 via RPC
>         DSA object GUID: 81a27497-bdfb-4977-9874-675bbfba490f
>         Last attempt @ Tue Jul  1 20:58:53 2014 CEST failed, result 
> 8442 (WERR_DS_DRA_INTERNAL_ERROR)
>         825 consecutive failure(s).
>         Last success @ Tue Jul  1 16:34:57 2014 CEST
>
> CN=Configuration,DC=samba,DC=domain,DC=com
>     Default-First-Site-Name\DC1 via RPC
>         DSA object GUID: 81a27497-bdfb-4977-9874-675bbfba490f
>         Last attempt @ Tue Jul  1 20:58:13 2014 CEST was successful
>         0 consecutive failure(s).
>         Last success @ Tue Jul  1 20:58:13 2014 CEST
>
> Google doesn't tell me much. Does anyone know WERR_DS_DRA_INTERNAL_ERROR?
>
> I rebooted the dc1, but that made no difference. Time is synched on 
> both dc's.
>
> Samba-tool dbcheck gives the same output on both dc's, and nothing out 
> of the ordinary. (it gives normalization errors, but these have been 
> discussed on list, and are considered harmless)
>
> Any cue's how to solve this? Stupid detail: I'm currently on holiday, 
> so much for a worry-free holiday....
>
> Hope for any insight.
>
> MJ