Strange Gateway Error with one APP Server (WSAETIMEDOUT)

Question

Hello Everyone,
We are experiencing a strange gateway timeout issue on one application server in our SAP environment.  This started happening after we applied a series of Microsoft patches and updated the SAP kernals from 179 (DB 172) to 185 (DB 182).
From this App server we are seeing all kinds of gateway errors, the following is a sample from the dev_rd file from the server in question:
  LOCATION    SAP-Gateway on host sappe1exp5 / sapgw00
  ERROR       partner 'sappe1exp4:sapgw00' not reached
  TIME        Sat May 30 23:40:54 2009
  RELEASE     700
  COMPONENT   NI (network interface)
  VERSION     38
  RC          -10
  MODULE      nixxi.cpp
  LINE        2823
  DETAIL      NiPConnect2
  SYSTEM CALL connect
  ERRNO       10060
  ERRNO TEXT  WSAETIMEDOUT: Connection timed out
  COUNTER     67
GwTraceHdlInfo:
HANDLE   = 6   
TIME     = Sat May 30 23:40:33 2009
SOCKET   = 468
STAT     = NI_CONN_WAIT
TYPE     = STREAM IPv4
OUT      = 0 messages 0 bytes
IN       = 0 messages 0 bytes
LOCAL    = 0.0.0.0:2604
REMOTE   = -
OPTIONS  = BUFFERED
  ni hdl      = 6
  type        = CLIENT
  net_stat    = CONNECT_TO_REM_GW_PENDING
  hostaddr    = 10.10.0.21
  opcode      = NORMAL_CLIENT
  conn opcode = REMOTE_GATEWAY
  conn vers   = 105
  index       = 59
  data        = 0000000000000000
  ext_info    = 000000000333DCA0
  offset      = 0
  rest_len    = 0
  snc_forced  = 0
remote gateway infos:
  hostname =
  hostaddr = 10.10.0.21
  service  = sapgw00
  tpname   = sapdp00
Sat May 30 23:46:17 2009
***LOG Q0I=> NiPConnect2: connect (10060: WSAETIMEDOUT: Connection timed out) [nixxi.cpp 2823]
ERROR => NiPConnect2: SiPeekPendConn failed for hdl 6 / sock 428
    (SI_ECONN_REFUSE/10060; I4; ST; 10.10.0.21:3300) [nixxi.cpp    2823]
I have done alot of searching and I have tried the following things:
1) Checked the services file of this machine, they line up with all other systems
2) Attempted to telnet to other App servers on 3200, 3300, 3600 with no success
3) tried to connect via SM59 to another App server with no success
4) Attempted to do all of these things in reverse from the another app server with the same results.
This is only happening with one app server in our landscape and this app server has no issues connecting to the CI.
Any ideas on where I can look further?
Edited by: James Wright on Jun 5, 2009 10:10 PM

Answer 1

Hi,
on sappe1exp5  try the following command in a dos prompt:
telnet sappe1exp4 sapgw00
this should also time out i.e. hang for a certain amount of time (~50 seconds) and then terminate with error.
It simply means that the IP address 10.10.0.21 is not reachable from your client PC.
Maybe you changed some IP addresses during patches. Usually there is an RFC destination that is not functioning properly.
Best regards
Carl Philipp