cancel
Showing results for 
Search instead for 
Did you mean: 

Error in ERP's cache (SXI_CACHE)

Former Member
0 Kudos

Hello, All!

We try to customize an integration between SAP ERP 6.0 (NW7.31, SP08) and SAP PI 7.11 (SP12).

Name them ERP and PI as appropriate.

All necessary RFC-connections are working fine and SPROXY have generate appropriate Services Interfaces successfully.

I've made a simple integration scenario in PI. And when I to send the message from ERP, I see always the same error in ERP's SMB_MONI:


SAP:Code area="PARSING">GENERAK<SAP:Code>

Couln't retrieve inbound binding for the given P/S/A values ...

I've found out the solution = refresh CPA cache in ERP-side.

I'm going to t-code SXI_CACHE and see the next Errors on the first page:

Unable to refresh cache contents.

Errpr during last attempt to refresh cache.


In this transaction I press to F7 for test settings and see:

Cache Status = Inconsistent

Error during last update = http_body_size = 0



Thought I haven't any errors at PI-side. I can execute CPA FULL Refresh. I've restart both systems.


Please help me to update the cache.

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hallo Guys!

Have resolved.

It was not a error. It's ERP's normal behavior.

See Functional Unit Configuration UI scenarios.

It makes all necessaries RFC-connections, users & etc.

Have a fun.

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

Good News, please close the thread and asign point if you consider it

Former Member
0 Kudos

Hi Andrei,

Could you please elaborate on the issue.

Because, we are facing the issue in PI.

The link you provided in the post doesn't work. Did you run Functional UI scenarios in PI?


Appreciate your help here.

Thanks and Regards,

Satish

Former Member
0 Kudos

At first - "PI Self Test for NetWeaver"

Later "Functional Unit: SAP NetWeaver Process Integration (PI)".

Answers (2)

Answers (2)

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

Hi,

please execute the transaction SLD_CHECK and put an screenshot.
this errors relates with user that are bloqued like PIAPPLUSER,PIDIRUSER,PIISUSER

ALso check RFC destination TCP/IP LCRSAPRFC and SAPSLDAPI.

Let me know.

Former Member
0 Kudos

Hi,

I have executed SLDCHECK and it launched the SLD page successfully and all the RFCs are working fine. I am afraid i cannot post the screenshot here.

I have also checked SLDAPICUST and it works fine too.

I was able to successfully execute CPA Cache refresh on JAVA side.

I am skeptical about status of ABAP Cache in SXI_CACHE being Inconsistent.

Please let me know if i have missed any other config.

Thanks and Regards,

Saitsh

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

could you execute a CPACache <server:host>/CPACache/refresh?mode=full

let me know

I assume that a service has dump or a user is blocked.

lets see

Former Member
0 Kudos

Yes. I did execute the Full Cache refresh and it is done.

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

excuse me, did you execute the sldcheck in the ECC or PI system?

Former Member
0 Kudos

I have executed in PI.

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

if the error is in the ERP do the SLDCHECK in the ERP

allamudi_loordh
Active Participant
0 Kudos

Hi Andrei.

check this user PIDIRUSER whether it is locked or not. or due complete cache refresh with below url

http://host:port/CPACache/refresh?mode=full

regards,

Loordh.

Former Member
0 Kudos

Hi, Allamudi!

I've executed complete cache refresh successfully via this link early. PIDIRUSER is valid and can do this refresh. Tests in SXMB_ADM executed successfully too.

I haven't any errors at PI-side and SXI_CHACHE in PI presents correct data, but not in ERP...

But PI-Side haven't any messages in SXMB_MONI.

Former Member
0 Kudos

Allamudi,

At PI side, in NWA Java monitoring I can see the suspicious Error:

(In time when I try to update cache in ERP)

Cannot process an HTTP request to servlet [CacheRefresh] in [dir] web application.

allamudi_loordh
Active Participant
0 Kudos

Hi Andrei,

did you create all the objects in ID. and verify them in sxi_cache.  and also  check this destination working or not INTEGRATION_DIRECTORY_HMI

regards,

Loordh.

Former Member
0 Kudos

Allamudi,

The conectiong is working fine: HTTP Status 500 [Internal Server Error].

I've read the userguide where this situation named as Ok.

I feel that it's an error with Servlet CacheRefresh and his application dir.

I've opened this in NWA. What I must to do there?

Former Member
0 Kudos

Additional screenshot.

allamudi_loordh
Active Participant
0 Kudos

Hi Andrei,

you should get 200 status ok in RFC destination ping.Check this user    PIISUSER

regards,

loordh.

Former Member
0 Kudos


Allamudi,

I don't think so.

In XI's Configuration Guide I can see some text: Http error 500 is normal state of communication/
.

Yes, it's so strange and many people ask about it in SCN.SAP.

Anyway I see the same situation with the connection test - in PI and ERP side, thought in PI side it was generated automatically.

PIISUSER (in PI) is ok and he has SAP_ALL permissions.

allamudi_loordh
Active Participant
0 Kudos

check what user is used in that rfc destination.and it should be 200.

Former Member
0 Kudos

Allamudi,

User PIISUSER is in the rfc conn.

But I don't understand why my return code is "HTTP 500" is error?

I've read the forum scn.sap.com/thread/1671367 and many-many others threads.

Where was discussed 500 code error.

Tell me, what do you think about servlet CacheRefresh? Maybe the error in that?

former_member184720
Active Contributor
0 Kudos

Hi Andrei - I don't think it's an issue your ECC cache as the objects will never exists in ECC. You have to make sure that PI cache is update.

You can try the below approach - for proxy interfaces are you using integrated configurations?(soap adapter) or Proxy ?(XI adapter)

If it is Soap Adapter -> test it from Soap UI

If it is Proxy channel -> test it from RWB

Another check ->

If you have Ico then please check if the below Software component version is assigned for the sender interface.

Former Member
0 Kudos

Hi, Hareesh!

I don't use the SOAP channel as yet. But I can't see the XI-sender channel in RWB.

It's so strange too.

The cache update via host:port/CPACache/refresh?mode=full is OK.

When I try to update cache in ERP, I immediately see always the same error in PI's NWA JAVA-logs:

Cannot process an HTTP request to servlet [CacheRefresh in [dir] web application.

Former Member
0 Kudos

Tcode SLG1 shows in time of CacheRefresh in ERP.

What is "http_body_size = 0" error? I can't find out in SAPNotes...

Former Member
0 Kudos

Haaresh,

When I enter to host:port/dir/CacheRefresh get an Error 500 Internal Error.

I don't truth that it's OK.

In another PI-host return only one word "OK" without any errors (and http code 200).

From there cache is refresh successfully.

Where I must customize PI side?

Re-deploy java-applications or upgrade pi from 7.11 to 7.31?

former_member184720
Active Contributor
0 Kudos

Hi Andrei - Are you getting this error when you test the RFC destination  "INTEGRATION_DIRECTORY_HMI" in PI system?

Howvever Go to SXI_CACHE and hit F7 and share the screenshot of it (in PI system)

Former Member
0 Kudos

Good day, Haaresh!

I've taken the screenshot from PI.

Moreover I installed the new package for SAP_BASIS (mentioned in snote 1933454).

But this didn't help me.((

Message was edited by: Andrei Sokolov

former_member184720
Active Contributor
0 Kudos

Your SXI_Cache seems to be fine but SLG1 is giving some information (reporting error in cache refresh to directory - 808)

However from the PI tools page

go to administration -> directory -> data cache

there try to refresh the cache and see if ithat works..

If it doesn't work try restarting the system

Former Member
0 Kudos

Good day, Haaresh!

I've executed this operation. As I sad early, it works:

former_member184720
Active Contributor
0 Kudos

Hi Andrei - Then PI system is good.

Can you confirm one thing - Did you try running the interface end-end?

Incase you are worried about the The reason why your XI channel is not showing in CC monitoring then it is because "XI" and "HTTP" adapters resides on the Integration engine and hence you can not monitor them in AE.

I think i over looked you reply as i thought it is in PI system.. i just read it again..

"Tcode SLG1 shows in time of CacheRefresh in ERP.

What is "http_body_size = 0" error? I can't find out in SAPNotes"

I'm not sure if you can use it in ECC system as it is for XI directory cache.

Incase you get the error Couln't retrieve inbound binding for the given P/S/A values - make sure the object exists in Integration directory  for that combination(copy the values from ECC sxmb_moni and search)

Former Member
0 Kudos

Hi Andrei,

Did you find the resolution of the issue?

I am facing exact same issue in our Dev system.

Please share any knowledge on the issue.

Regards,

Satish