[Samba] smb connection closed

Jeremy Allison jra at samba.org
Wed Oct 24 19:30:07 UTC 2018


On Tue, Oct 23, 2018 at 05:04:22PM +0200, Sascha Wiechmann via samba wrote:
> Hi everybody,
> 
> May someone help me out on a short question, please? We own an ERP based on
> MSAccess, backend is on a samba share, openSuSe 42.3. The ERP is very
> sensitive on disconnection from those - also if it is just a reload of
> config.
> We did not changed anything within the last week, but on thusday/friday my
> collegues started to comlain about "database or network error" message of
> our ERP. I started to check network etc and increased loglevel on
> samba-machine.
> But, as really nothing changed in the whole month from my site, I did not
> think about a samba problem.
> 
> Today I checked the logs from one machine disconnecting quite often. My
> Question is : Do I understand well when I read the machine is closing the
> connection from client side?
> As you can see, it tries to get oplock (am I right?) and then starts to
> close all files and ends connection.
> 
> ............................
> 
> [2018/10/23 09:27:35.529661,  3]
> ../source3/smbd/smb2_read.c:413(smb2_read_complete)
>   smbd_smb2_read: fnum 591422384, file Navidata/Common/Cash.mdb,
> length=16384 offset=85655552 read=16384
> [2018/10/23 09:27:35.531806,  3] ../source3/lib/util.c:1183(fcntl_getlock)
>   fcntl_getlock: fd 45 is returned info 2 pid 0
> [2018/10/23 09:27:35.531860,  3]
> ../source3/smbd/smb2_read.c:413(smb2_read_complete)
>   smbd_smb2_read: fnum 591422384, file Navidata/Common/Cash.mdb,
> length=16384 offset=85671936 read=16384
> [2018/10/23 09:27:35.533803,  3] ../source3/lib/util.c:1183(fcntl_getlock)
>   fcntl_getlock: fd 45 is returned info 2 pid 0
> 
> ................
>   samba1user closed file Navidata/Exchange/CustC/FUGermany.ldb (numopen=32)
> NT_STATUS_OK
> [2018/10/23 09:28:46.678381,  2]
> ../source3/smbd/close.c:788(close_normal_file)
>   samba1user closed file Navidata/Common/AllCustomers.ldb (numopen=31)
> NT_STATUS_OK
> [2018/10/23 09:28:46.678534,  2]
> ../source3/smbd/close.c:788(close_normal_file)
>   samba1user closed file Navidata/Common/Workbench.ldb (numopen=30)
> NT_STATUS_OK
> [2018/10/23 09:28:46.678696,  2]
> ../source3/smbd/close.c:788(close_normal_file)
>   samba1user closed file Navidata/Retoure.ldb (numopen=29) NT_STATUS_OK
> [2018/10/23 09:28:46.679274,  2]
> ../source3/smbd/close.c:788(close_normal_file)
> 
> ......
> 
> [2018/10/23 09:28:46.685862,  2] ../source3/smbd/service.c:1098(close_cnum)
>   escjwaw7p100 (ipv4:192.168.0.100:50105) closed connection to service
> navidataesc
> [2018/10/23 09:28:46.687056,  3]
> ../source3/smbd/server_exit.c:242(exit_server_common)
>   Server exit (NT_STATUS_CONNECTION_RESET)
>  .................
> 
> And is there a hint for the NT_STATUS_OBJECT_NAME_NOT_FOUND error? I googled
> a lot but did not find anything helpful - hoped I will find a debugging wiki
> :)

No, NT_STATUS_OBJECT_NAME_NOT_FOUND is a common return error in
normal use, whenever the client looks for a file that isn't there.

The key is in the:

 [2018/10/23 09:28:46.687056,  3]
 ../source3/smbd/server_exit.c:242(exit_server_common)
  Server exit (NT_STATUS_CONNECTION_RESET)

message which normally means the client reset/closed the TCP
connection.



More information about the samba mailing list