HTTPS Session can not be terminated in SM04

[j_bayrhammer]

Hello,

user is logged off from system, only one session can be seen in SM04.

It's a HTTPS Session, application info is:

R=7 P=/sap/bc/nwbc/nwbc_ext2int

User is logged off from system, no gui, no nwbc session is open.

If I try to log off user (via menu user - log off user - systemwide) session is still there.

If I try to get session information via doubleclick systems gives message "No sessions available; check if user is already logged off".

Sometimes I can see a process in SM50 with that user executing a program SAPMHTTP.

If I try to catch this program and stop it or kill the process, the same user with the same program returns after a few seconds.

Additionally user causes dumps in syslog every four seconds:

Signal 11 received by operating system

Stop Workp. 21, PID 6357138

Start Workp. 21, 58 times since system startup. PID 7 976492

PID is switching.

How can we stop this?

Regards,

Julia

[mercedes_fernandez]

Hello Julia,

regarding the HTTPS sessions in SM04 see SAP note "2211359 : SM04 displays HTTPS-sessions after logging them off"

To find out the reason for signal 11 is needed to identify the function where it happens.
KBA "2162344 : SYSTEM_CORE_DUMP : how to analyze it" describes how to find out that function.

hope it helps,
Mercedes

[j_bayrhammer]

Hello Mercedes,

with second note I think you meant note number 2162346 instead of 2162344.

In note 2211259 the two parameters rdisp/security_context_enhanced_trace and rdisp/security_context_array_termination are mentioned, but no further Information on how to use it. Can you please give some more Information about these parameters?

Regards

Julia

[mercedes_fernandez]

Hello Julia,

yes, I meant 2162346

regarding the other one, you need to check if the field "security_context_reference" in SM04 --> Technical information are the same for all the pending sessions.
Is that the scenario?

thanks,
Mercedes

[j_bayrhammer]

Hello Mercedes,

I can't compare if this is the same value as other pending sessions, there is only this one pending session.

Regards,

Julia

[mercedes_fernandez]

Hello Julia,

is it always the same user and url?
when do you get the signal 11, when trying to remove it?
Which kernel release is the system running in?

thanks
Mercedes

[j_bayrhammer]

Hello Mercedes,

it's just one user and one session.

We have this problem since yesterday with only one user with only one session.

Signal 11 is created "itself" - nobody tries anything. Every four seconds there is a new entry in SM21 with signal 11 error.

It's kernel 742 with patch level 122.

RZ11 does not know parameters rdisp/security_context_enhanced_trace and rdisp/security_context_array_termination, so I can't find any information in SAP about it.

Regards,

Julia

[mercedes_fernandez]

Hello Julia,

as the system is in kernel 742 that note / parameters are not relevant.
SAP note 2211359 is relevant for kernels 721/722

Are those signal 11 related to that entry?

If you think they are,then you'd need to check the C-stack and see which function triggers it.
see the note I mentioned before ( 2162346)

Would it be possible you copy the Technical data here for that pending session?

thanks,
Mercedes

[Former Member]

Hi Julia,

Please check the work process trace file to get more details. Activate the trace on the user too.

If signal 11 is received than there has to be core file generated at operating system level in work directory may be.

Check below

2162346 - SYSTEM_CORE_DUMP: How to analyse it

1964673 - C-Call stack analysis

Extracting the C-Stack from core files (UNIX/Linux/AIX):

Extracting the C-Stack from core files (UNIX and Linux) - Application Server Infrastructure - SCN Wi...

Execute the command to see whom, why n when it occurred?

$ errpt -a output

with this command to check who generated the file with

$ /usr/lib/ras/check_core core | tail -1

Regards,

Prithviraj.

[j_bayrhammer]

Hello,

thank you for your help.

I was not in office the last days - System was restarted a few times in meantime.

Error is now "disappeared".

I will keep that in mind and update this thread if problem occures again.

Happy new year & regards,

Julia

[former_member185239]

Hi Julia,

Follow the below steps

- Logon to the url http://<server>:<port>/nwa/app-resources

- Look for the SAPFactory

- check the Connection Lifetime value to 10 sec.

and then go through the sapnotes

http://service.sap.com/sap/support/notes/1805710

1970650 - How to... Manage Connection Pooling

With Regards

Ashutosh Chaturvedi

[j_bayrhammer]

Hello Ashutosh,

it is not a double stack system, it is an abap only System.

So there is no NWA.

Regards,

Julia

[former_member185239]

Dear Julia,

I believe that users are accessing the backend ABAP system from Portal system or HTML GUI.

If the users are accessing the backend system from a portal system then you have to maintain the SAPFactory parameter on portal side.

Also activate the logoff SICF services on the ABAP system.

Also check the sapnotes suggested by

With Regards

Ashutosh Chaturvedi

[j_bayrhammer]

Hello Ashutosh,

user is accessing ERP system directly through nwbc gui. No Portal or HTML GUI between.

With logoff SICF parameter you mean this one:

sap - public - bc - icf - logoff?

Why this problem is only with one user?

Regards,
Julia