Operating system call connect failed | SM21

In SM21 system messages there are several entries for Operating system call connect failed as below

• Operating system call connect failed (error no. 113)
• Operating system call connect failed (error no. 115)

• 

You see the Type as IC. Type IC means the message is coming from ICM. So we should check ICM traces from SMICM or from OS level, dev_icm file in work directory (/usr/sap/<SID>/<instanceNN>/work) of the same instance that’s shown in SM21 under Instance column. Below is the logs found in ICM trace for error 113 EHOSTUNREACH. In this case, ABAP system tries connecting to Open text server but the connection is not established. Here the server/IP is not resolving or blocked for port 8080. We can confirm this by testing telnet from the instance.
telnet 192.160.128.163 8080

dev_icm
[Thr 140231990015744] Fri Jun 23 06:04:54:019 2017
[Thr 140231990015744] ***LOG Q0I=> NiPConnect2: 192.160.128.163:8080: connect (113: No route to host) [/bas/753_REL/src/base/ni/nixxi.cpp 3450]
[Thr 140231990015744] *** ERROR => NiPConnect2: SiPeekPendConn failed for hdl 110/sock 34
    (SI_ECONN_REFUSE/113; I4; ST; 192.160.128.163:8080) [nixxi.cpp    3450]
[Thr 140231990015744] *** WARNING => Connect to chncs1ot01.sap.m1bc.com:8080 failed: NIECONN_REFUSED(-10) {000200cc} {root-id=00163E5EBAC91EE9A5B4D69F2722B944} [icxxconn.c 2604]
[Thr 140231990015744]                INTERNAL T32_U1299_M0, 000, SAPSYS, , time=06:04:53, W26, program=SAPMSSY8, high priority, memory=0, tasks=1, appl info=, tcode=
[Thr 140231990015744]                role: Client, protocol: HTTP

Similary for error 115 (EINPROGRESS). It tried connecting to the server but it takes long time to get response.

dev_icm 
[Thr 140231991072512] Fri Jun 23 06:39:49:619 2017      
[Thr 140231991072512] ***LOG Q0I=> NiBufIConnect: connection pending after 5000ms: connect (115: Operation now in progress) [/bas/753_REL/src/base/ni/nibuf.cpp 4568]  
[Thr 140231991072512] *** ERROR => NiBufIConnect: non-buffered connect pending after 5000ms (hdl 105; 192.160.128.163:http-alt) [nibuf.cpp    4579]  
[Thr 140231991072512] *** WARNING => Connect to chncs1ot01.sap.m1bc.com:8080 failed: NIECONN_REFUSED(-10) {0002027b} {root-id=00163E5EBAC91EE9A5B573158A0BBAE4} [icxxconn.c 2604]  
[Thr 140231991072512]                INTERNAL T26_U10373_M0, 000, SAPSYS, , time=06:39:44, W34, program=SAPMSSY8, high priority, memory=0, tasks=1, appl info=, tcode=  [Thr 140231991072512]                role: Client, protocol: HTTP

In this case, there is an activity on the server chncs1ot01.sap.m1bc.com which is the reason for these syslog messages. So one needs to check the logs first which will help in narrowing down the analysis.

There may be different reasons for such errors. In few cases, it may be due SMTP server as well.

SAP KBA “2582112 – ERROR => NiBufIConnect: non-buffered connect pending after ms“ tells how to analyse the issue.

Also there are some known issues as described in KBAs

2147794 – “Operating system call connect failed” errors in System Log after removing a server from landscape and

2727608 – Obsolete repository in KPRo.

In case of obsolete repositories, the errors would be like below as described in KBA.
*** WARNING => Connect to 10.26.24.44:1090 failed: NIECONN_REFUSED(-10)
*** WARNING => Connect to P25214:1090 failed: NIEHOST_UNKNOWN(-2)
*** WARNING => Connect to http:80 failed: NIEHOST_UNKNOWN(-2)
*** WARNING => Connect to pwdf2625:1090 failed: NIEHOST_UNKNOWN(-2)
*** WARNING => Connect to arcdev.wdf.sap.corp:1080 failed: NIEHOST_UNKNOWN(-2)
*** WARNING => Connect to hs2086.wdf.sap.corp:2000 failed: NIEHOST_UNKNOWN(-2)
*** WARNING => Connect to dwdf010.wdf.sap-ag.de:1090 failed: NIEHOST_UNKNOWN(-2)
That’s if you see one of these hostname/IP and port in the error log, then check whether KBA 2727608 applies to you.