WSAENOBUFS error

[Former Member]

Hi experts,

I am getting the error below on both my Dev and Prod servers.

The SAP MMC will still show everything as green but the SAP system will be inaccessible.

This error was found in one of the logs in the work directory.

Sun Jan 26 09:33:57 2014

A  AppServerTimeSync, getDBsynctime_strict():

A    returning successfully with

A    Localtime: 1390700037, DBTime: 1390700037.

A    Duration of DB calls(microsecs): DBTime1: 1063, DBTime2: 304.

A  AppServerTimeSync, zdate_gettime():

A    New bias computed: 0.

M  ***LOG Q0I=> NiIBindSocket: bind (10055: WSAENOBUFS: No buffer space available) [D:/depot/bas/740_REL/src/base/ni/nixxi.cpp 3759]

M  *** ERROR => NiIBindSocket: SiBind failed for hdl 416/sock 1492

    (SI_ESYSCALL/10055; I4; DG; 127.0.0.1:0) [nixxi.cpp    3759]

M  {root-id=32444635423834353244463542383435}_{conn-id=00000000000000000000000000000000}_0

M  *** ERROR => RqQQueueGetSelectableHandle: NiDgHdlBindName failed (rc=-1) [dpQueue.c    2679]

M  *** ERROR => ThRqGetSessionQueueSelectHandle: RqQQueueGetSelectableHandle failed (-1) [thRequest.c  5159]

M  ***LOG Q0I=> NiIBindSocket: bind (10055: WSAENOBUFS: No buffer space available) [D:/depot/bas/740_REL/src/base/ni/nixxi.cpp 3759]

M  *** ERROR => NiIBindSocket: SiBind failed for hdl 417/sock 1492

    (SI_ESYSCALL/10055; I4; DG; 127.0.0.1:0) [nixxi.cpp    3759]

M  *** ERROR => RqQQueueGetSelectableHandle: NiDgHdlBindName failed (rc=-1) [dpQueue.c    2679]

M  *** ERROR => ThRqGetSessionQueueSelectHandle: RqQQueueGetSelectableHandle failed (-1) [thRequest.c  5159]

M  *** ERROR => ThRqCheckQueuesCreateSelectSet: ThRqGetSessionQueueSelectHandle failed [thRequest.c  1261]

M  *** ERROR => ThRqCheckQueues: ThRqCheckQueuesCreateSelectSet failed (rc=16789512) [thRequest.c  1649]

M  in_ThErrHandle: 1

M  *** ERROR => ThRqCheckQueues: Fatal error (step TH_WORK_NOSYNC, thRc ERROR-CORE-RECEIVE_DATA, action RESTART_WP, level 1) [thxxhead.c   2158]

M  ThIErrHandle: set thRc ERROR-CORE-RECEIVE_DATA for T27_U22604_M0

M  ThResetEmMagic: reset em magic for T27_U22604_M0_I0

Both servers are on Windows Server 2012 and on 64 bit. It cannot be due to a memory issue as my Prod server has more memory yet is facing the issue as well.

I have also made the registry changes recommended in note 990538 for both MaxUserPort and TcpTimedWaitDelay but it doesn't seem to have any effect.

In every case, we have to restart the server for the system to be accessible again.

Thanks and regards.

[Former Member]

Hi Gregory,

I have some questions for you to answer:

1. which kernel version you are running on the ABAP Server?
2. did you capture an overview about the used sockets at the time, the problem shows up?
   if not: please run from an elevated cmd.exe or powershell:
      netstat  -ano  > OpenPorts.txt
      netsh interface ipv4 show dynamicportrange protocol=tcp > PortRange.txt
      netsh interface ipv4 show dynamicportrange protocol=udp >> PortRange.txt

please provide the contents of both files (the first has a lot of lines in this context so zipped upload is better).

kind regards

Peter

[Former Member]

Hi Peter,

My kernel is release 740 level 37. This error usually happens over the weekend and the system will be accessible when we come back to check it on Monday.

I've run the commands and exported the results. I hope you don't mind providing an address where I can send them to.

Thanks and regards.

[Former Member]

Gregory,

please update your kernel to level 46 or higher.

( 0.046) Memory leak in queue handling (note 1947355) describes the details.

kind regards

Peter

[Former Member]

Hi Peter,

For kernel 740, level 37 is still the highest that I can find in the SMP. Will a higher level be released soon?

Also, I can't view that SAP note as it's saying that it has not been released yet.

Thanks and regards.

[Former Member]

Hi Gregory,

as far as I can see (as I am an SAP Employee, I have a different view on the service market place) the latests SAPEXE.SAR is of level 37 (as you already mentioned).

The latest dw.sar is of patchlevel 52.

dw.sar contains the abap-server and depended modules.

You get always the current version when applying latest sapexe.sar and latest dw.sar.

this is also described in SAP note 19466.

The above mentioned note was updated yesterday to contain also the error message.

Because it's master language is in German it needs still to be translated, which should be finished in the next days.

Until then please refer to the previous version I attached (rename 0001947355.pdf.txt to 0001947355.pdf after downloading)

kind regards

Peter

0001947355.pdf.txt.zip

[Former Member]

HI Peter,

I've updated the dw.sar to level 52 and the error stopped for awhile but it has just reoccurred again. The log entry is as below.

Mon Feb 10 03:47:41 2014

M  ***LOG Q0I=> NiIBindSocket: bind (10055: WSAENOBUFS: No buffer space available) [D:/depot/bas/740_REL/src/base/ni/nixxi.cpp 3759]

M  *** ERROR => NiIBindSocket: SiBind failed for hdl 323/sock 1384

    (SI_ESYSCALL/10055; I4; DG; 127.0.0.1:0) [nixxi.cpp    3759]

M  {root-id=00155D0001261ED3A4B86113BCB9F410}_{conn-id=00000000000000000000000000000000}_0

M  *** ERROR => RqQQueueGetSelectableHandle: NiDgHdlBindName failed (rc=-1) [dpQueue.c    2988]

M  *** ERROR => ThRqGetSessionQueueSelectHandle: RqQQueueGetSelectableHandle failed (-1) [thRequest.c  5256]

M  ***LOG Q0I=> NiIBindSocket: bind (10055: WSAENOBUFS: No buffer space available) [D:/depot/bas/740_REL/src/base/ni/nixxi.cpp 3759]

M  *** ERROR => NiIBindSocket: SiBind failed for hdl 324/sock 1384

    (SI_ESYSCALL/10055; I4; DG; 127.0.0.1:0) [nixxi.cpp    3759]

M  *** ERROR => RqQQueueGetSelectableHandle: NiDgHdlBindName failed (rc=-1) [dpQueue.c    2988]

M  *** ERROR => ThRqGetSessionQueueSelectHandle: RqQQueueGetSelectableHandle failed (-1) [thRequest.c  5256]

M  *** ERROR => ThRqCheckQueuesCreateSelectSet: ThRqGetSessionQueueSelectHandle failed [thRequest.c  1282]

M  *** ERROR => ThRqCheckQueues: ThRqCheckQueuesCreateSelectSet failed (rc=16789512) [thRequest.c  1670]

M  in_ThErrHandle: 1

M  *** ERROR => ThRqCheckQueues: Fatal error (step TH_WORK_NOSYNC, thRc ERROR-CORE-RECEIVE_DATA, action RESTART_WP, level 1) [thxxhead.c   2243]

M  ThIErrHandle: set thRc ERROR-CORE-RECEIVE_DATA for T17_U19678_M0

Is there another part of the kernel that needs to be updated as well?

Thanks and best regards.

[Former Member]

Hi Gregory,

OK, it looks like this is a new one:
please open a message at SAP support!

kind regards

Peter

[Former Member]

Hi all,

The boot.ini step is only relevant for 32 bit systems. It is 64-bit in our case. The server restart will solve the issue but it will then reoccur on a weekly basis.

The DB is DB2 10.05.0001 on Windows Server 2012.

Thanks and regards.

[Sriram2009]

Hi Lim

During this Error have you check the ST02 Buffer parameters are full (More Swaps) ? Kindly check this SAP Notes 129813 - Windows: Problems due to address space fragmentation

and also have you follow the SAP Note 88416 - Zero administration memory management for the ABAP server

Regards

Sriram

[Former Member]

Hi all,

I've already mentioned in my post that I've checked out that note and set all the entries in the registry for both MaxUserPort (65534) and TcpTimedWaitDelay (5) but it is still the same.

Is there anything else to check?

Thanks and regards.

[former_member182034]

hi Gregory,

Please follow the instruction from below link. it might be helpful

WSAENOBUFS: No buffer space available in SAP

Regards,

[Sriram2009]

Hi

OS /DB version ?

Regards

Ram

[former_member182034]

hi Gregory.

Please go through below link. It might be helpful

'WSAENOBUFS (10055)'

Regards,