on 09-23-2016 11:43 AM
Dear Experts,
We're migrating SAP system from Oracle to HANA using SWPM. Export was successful but with import we're struck with the error in Start Instance step of execution phase using SWPM (software update maanger).
From dev_disp log file, we found the following error:
Error NiPConnect2: SiPeekPendConn failed for hdl 17/sock 10
Error MsIAttachEx: NiBufconnect to hostname/3901 failed (rc=NIECONN_REFUSED) [msxxi.c 878]
You're ideas are appreciated.
Thanks & best regards,
Sreenu
Check if the /etc/hosts and etc/services files are set up correctly.
Also review the profile parameters related to message server.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I was wondering, if the system is connected to your company domain server or not. In this given scenario, I would have tried flushdns & then telnet to that server from another system. If it was windows, then I would have suggested to install called Microsoft Loopback Adapter. Anyways because the system here instead of takingthe system IP 1st, its taking the loopback IP 1st then, the system IP
I was also thinking, Is there any saphostcntrl/saphostagent version included?
Thanks,
Kaushik
Assisting you is difficult as you are not supplying the full trace logs instead screenshots of parts of the trace files. You need to supply the full trace files (attach them) from the work directory.
So far it appears to me a issue related to firewall.
Look into this note and follow the instructions.
Do a niping test to the MS port on the host where the ASCS is running and check the output.
Get someone from the network team and review the firewall settings and the ports.
Hi,
As per suggestions of all other colleagues, please provide full logs details instead of only screen shot. Also could you please stop ASCS (message server) [as no other instance are running] and try to take a reboot of the host ? Sometime due to heap memory issue also dispatcher doesn't get started and causes issue while starting up the system. After reboot , try to start up the instances manually and then run SWPM. If error still persists, please provide latest logs of dev_ms & dev_disp logs . also please check if it's generating dev_w** logs under work directory of DVEBMGS** directory , if yes, provide those information as well .
Regards,
Nilutpal.
Hello Sreenu,
Anil has already asked you to check the logs. Could you also run the following to check that all processes are running: sapcontrol -nr 01 -function GetProcessList
It looks to me as though you selected "Create Message Server Access Control List" in the define parameters step of SWPM. That being the case you will have to create entries in the ms_acl_info file. Please consult notes 1495075 and 826779.
KR,
Amerjit
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Sreenu,
Then I would edit the ms_acl_info file as referenced in the notes I mentioned.
I would take a copy of the existing the file first, delete the existing entries and add HOST = * to the file. Post migration you must revisit the file and restrict to specific IPs and Servers (proper security).
After the changes, restart your ASCS (just to be 100% sure) and then try and continue with SWPM.
KR,
Amerjit
Hi,
Dispatcher trace file shows:
NiPConnect2: SiPeekPendConn failed for hdl 17/sock 10
MsIAttachEx: NiBufconnect to hostname/3901 failed (rc=NIECONN_REFUSED) [msxxi.c 878]
It seems that it's unable to connect to Message Server.
It may happen that Message Server is not running (but you provided 'GetProcessList' output, showing
that MS is running). Or MS is not listening on that port, you should check Ms profile parameters to ensure it's listening on that port.
Is eventually connection to MS established?
It would be useful if you could attach into this thread MS trace file as well as Disp trace file.
Regards,
Raquel
Hello,
Have a look to dev_ms logs and check if require services of message server has been started or not.
Regards,
Anil Bhandary
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thanks.
In that case, can you check the dev_disp on which port it is trying to check the message server host. Also same information should be in instance profile. Based on profile detail try to telnet hostname on port 3901 and check if get the result as expected.
Also , is your CI and CS is running on same server ?
Regards,
Anil Bhandary
Hallo Anil,
I see that the dev_disp using correct port 3901 for checking the message server host. And also in default profile, same port 3901 is maintained for parmater rdisp/msserv_internal.
We're getting result as expected using telnet. CS and ASCS are running on the same server.
Any other ideas?
Thanks,
Sreenu
User | Count |
---|---|
93 | |
10 | |
10 | |
9 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.