SOLMAN7.1-Installation issue-Connect to message server failed: MSENILAYER

[Former Member]

Hi all,

we've faced this error when installing the central instance of a distributed system of Solution Manager in Enterprise Linux 5 with an Oracle DB.

To give the full context, we installed first ERP6.0 EhP5 (ABAP stack only) under the same platform (with a Solution Manager key provided externally), but we found out that the Telco industry solution is not installed up to EhP5 level, and we're missing funcionality, so we decided to install Solution Manager for the upgrade of the corresponding software packages of the ERP.

We installed the SOLMAN Globahost and the DB without any issue, re-using the existing ERP DB host for SOLMAN as well - Multiple Components on Single Database -. The deployment of the SOLMAN schemas worked fine.

However, we're now installing the Central Instance, and BEFORE the actual installation process really starts, we come across with the following error message:

INFO 2012-02-21 17:03:06.535

Execute step askJavaMs of component |NW_Doublestack_CI|ind|ind|ind|ind|0|0|NW_GetSidFromProfilesPartial|ind|ind|ind|ind|4|0|NW_checkMsgServer|ind|ind|ind|ind|2|0

WARNING 2012-02-21 17:03:06.893

Connect to message server (saperp6app/3902) failed: MSENILAYER.

And the installation cannot continue. I've been checking similar issues but found no clue, we've also tried to change the port number but nothing changed. Any idea of what can be?

Thanks in advance.

Raúl.

[michael_ruth3]

Looks like someone tried to do a dual stack install

[Former Member]

Hi Michael,

Not sure to have understood your answer. As far as I know, SAP Solution Manager is a dual-stack application, but the ERP we're trying to upgrade is an ABAP only distributed system, as we read in note 855534 - Dual Stack and SAP Business Suite 7, that dual-stack systems are no longer supported.

But the only options we have for Solution Manager is dual stack.

Is there something wrong with this setup?

Regards. Raúl.

[Former Member]

Hi,

Could you please see if the port 3902 is reserved in /etc/services ?

sapms<SID> 3902

Br,

Venky

[Former Member]

Hi Venky,

Thanks for your answer. The entries I see in /etc/services for that port are:

nimaux 3902/tcp # NIMsh Auxiliary Port

nimaux 3902/udp # NIMsh Auxiliary Port

As we did not that port listening, we understood that it was not active although it is reserved.

Anyway, we've just modified the profile settins for a port we don't find in /etc/services (3976)

SOL_SCS02_saperp6app:

#----

-

---

1. SAP Message Server for Java

#----

-

---

rdisp/msserv = 0

rdisp/msserv_internal = 3976

ms/standalone = 1

ms/server_port_0 = PROT=HTTP,PORT=81$$

DEFAULT.PFL:

#----

-

---

1. SAP Central Service Instance for J2EE

#----

-

---

j2ee/scs/host = saperp6app

j2ee/scs/system = 02

j2ee/ms/port = 3976

login/system_client = 001

SAPDBHOST = saperp6db

j2ee/dbtype = ora

j2ee/dbname = I02

j2ee/dbhost = saperp6db

DEFAULT.1.PFL (not really sure whtat this file is):

#----

-

---

1. SAP Central Service Instance for J2EE

#----

-

---

j2ee/scs/host = saperp6app

j2ee/scs/system = 02

j2ee/ms/port = 3976

login/system_client = 001

Any ideas?

Regards. Raúl.

[Former Member]

Hi Raul,

Force an entry into the /etc/services file as follows.

Assuming ERP as your SAP SID.

sapmsERP 3976

Please ensure the port is not being used by another running process.

Use netstat -taupen |grep <port> or lsof -i:<port>

Please provide us any other error messages in the logs.

Br,

Venky

[nirmal_konchada]

Hi,

Also make sure that you have used different Instance no.'s for all the instances on the host.

Regards,

Nirmal.K

[Former Member]

Hi Venky,

Thanks for your answer, we've tried what you suggested with no success, still the same error.

This is the line in /etc/services we've added:

# more /etc/services | grep 3976

sapmsSOL 3976/tcp

What puzzles me is the fact that we haven't even started the installation of the Central Instance, we're in the initial steps in sapinst where the program asks for location of software, JDK and stuff llike that. In other words? Should not sapinst install that message server in the next steps? I don't get the issue. When installing the ERP we did not have to enter anything manually, I guess the system did it:

sapmsI01 3600/tcp # SAP System Message Server Port

Errors I see in log files in /tmp/sapinst_instdir/SOLMAN71/SYSTEM/ORA/DISTRIBUTED/AS/CI/

*******************

dev_sap_kernel

*******************

-

-

---

trc file: "dev_sap_kernel", trc level: 1, release: "720"

-

-

---

Wed Feb 22 09:51:03 2012

***LOG Q0I=> NiPConnect2: 170.251.121.47:3976: connect (111: Connection refused)

*** ERROR => NiPConnect2: SiPeekPendConn failed for hdl 1/sock 11

(SI_ECONN_REFUSE/111; I4; ST; 170.251.121.47:3976)

*** ERROR => MsIAttachEx: NiBufConnect to saperp6app/3976 failed (rc=NIECONN_REFUSED)

*******************

sapinst_dev.log

*******************

TRACE 2012-02-22 09:51:03.164

NWSCSInstance.getMSPortInternal() done: rdisp/msserv_internal = 3976

WARNING 2012-02-22 09:51:03.165

CIaMsgClient::CIaMsgClient(saperp6app, 3976)

Connect to message server (saperp6app/3976) failed: MSENILAYER.

TRACE 2012-02-22 09:51:03.165

handleException

*************

sapinst.log

*************

INFO 2012-02-22 09:51:02.358

Execute step checkMsgServer of component |NW_Doublestack_CI|ind|ind|ind|ind|0|0|NW_GetSidFromProfilesPartial|ind|ind|ind|ind|4|0|NW_checkMsgServer|ind|ind|ind|ind|2|0

INFO 2012-02-22 09:51:02.701

Working directory changed to /tmp/sapinst_instdir/SOLMAN71/SYSTEM/ORA/DISTRIBUTED/AS/CI.

INFO 2012-02-22 09:51:02.707

Execute step askABAPMs of component |NW_Doublestack_CI|ind|ind|ind|ind|0|0|NW_GetSidFromProfilesPartial|ind|ind|ind|ind|4|0|NW_checkMsgServer|ind|ind|ind|ind|2|0

INFO 2012-02-22 09:51:02.755

Working directory changed to /tmp/sapinst_instdir/SOLMAN71/SYSTEM/ORA/DISTRIBUTED/AS/CI.

INFO 2012-02-22 09:51:02.765

Execute step askJavaMs of component |NW_Doublestack_CI|ind|ind|ind|ind|0|0|NW_GetSidFromProfilesPartial|ind|ind|ind|ind|4|0|NW_checkMsgServer|ind|ind|ind|ind|2|0

WARNING 2012-02-22 09:51:03.165

Connect to message server (saperp6app/3976) failed: MSENILAYER.

I see a couple of errors that occur before, and I don't see if they are related:

****************

in sapinst.log

****************

WARNING[E] 2012-02-22 09:47:44.564

FSL-02053 Read error in /proc/sys/fs/binfmt_misc/register.

WARNING[E] 2012-02-22 09:47:44.643

FSL-00009 System call failed. Error 13 (Permission denied) in execution of system call 'fopenU' with parameter (/proc/sys/net/ipv4/route/flush, r), line (494) in file (syxxcfst

rm2.cpp), stack trace: syxxsyfact.cpp: 352: CSyslibFactory::writeSupportInfo(const PMsgLogBook & pLogBook)

...

WARNING[E] 2012-02-22 09:47:44.643

FSL-02015 Node /proc/sys/net/ipv4/route/flush does not exist.

WARNING[E] 2012-02-22 09:47:44.960

FSL-00009 System call failed. Error 13 (Permission denied) in execution of system call 'fopenU' with parameter (/proc/sys/net/ipv6/route/flush, r), line (494) in file (syxxcfst

rm2.cpp), stack trace: syxxsyfact.cpp: 352: CSyslibFactory::writeSupportInfo(const PMsgLogBook & pLogBook)

...

ERROR 2012-02-22 09:50:52.802

FSL-03003 Unable to determine parameter "DNS domain name". DNS domain name is not configured in resolv.conf (UNIX) or the registry (Windows) and neither getaddrinfo nor gethost

byname could retrieve it.

This is probably related to the fact that we haven't entered any DNS domain name for SAP system used to calculate the FQDN, but I don't think that's mandatory.

We're stuck now, don't know what else to check.

Regards. Raúl.

[Former Member]

Hi Nirmal,

our instance number for solution manager SOL is 02, whereas for the ERP previously installed is 00.

I've also checked with the instructions provided by Venky that there is neither port 3902 (deftault) nor 3976 (set by us to test) are in use.

What I remember is that by mistake we started the installation of the globalhost re-using the name I01 (ERP system) for Solution Manager, but we decided after checking discussions that that is not an advisable practice, so we started over with name SOL. and generated the folowing in I01_SCS01_saperp6app and START_SCS01_saperp6app.

Could that have anything to do?

Regards. Raúl.

[Former Member]

Hi, the issue is solved.

For whatever reason the message server of the instance SOL was corrupted, probably because we had to uninstall and re-install the DB. Instead of re-installing the DB, probably we should have started over.

I've been able to uninstall the instance SOL and reinstall it, the same system number was assigned and the server started. I'm proceeding with the installation.

Thanks everyone for your answers.

Regards. Raúl.