SXI_CACHE FAILED

[Former Member]

Dear expert,

I have a problem with the cache Of PI.

I work with PI 7.1 EHP1 SPS3.

In fact, The transaction is Ko, but i check the configuration and she seems to be good.

I check all dest in SM59, SLDCHECK, exchange profile etc...

The error message is:

Unable to refresh cache contents

Error during last attempt to refresh cache

Details:

Incorrect XML display for AlertCategory

Error during deserialization

Data loss occurred when converting 4d8a61e5f1063132a5a7ea30214ae5c1

The problem is the same in Integration directory when i want to refresh an object of entegration engine (Abap cache)

Thanks by advance

Edited by: Nor Goudjil on Nov 3, 2009 10:48 PM

[Former Member]

Hi Guru/Nor,

We too were suffering from the symproms described in this thread, namely symptom 3:

The Integration Server ABAP Cache Refresh of a changelist containing Alert Categories fails with the following error: "Incorrect XML display for AlertCategory", "Error during deserialization", "Data loss occurred when converting <GUID>". This error can be found in the "Cache Monitoring" section of the "Runtime Workbench" for the "Integration Server (ABAP Cache)" cache instance problem section.

After applying note 1394710 via SNOTE and following all the manual steps, the problem is resolved.

Thanks for the advice.

Regards,

Gary

[Former Member]

very useful thread to resolve SXI_CACHE failed error after PI 7.1 Ehp1 upgrade

Thanks a lot

Edited by: Govind Prathi on Dec 9, 2009 10:00 PM

[Former Member]

We had the same issue.. there is note for it which is 1394710..we need to implement it..Then it worked fine..SAP also suggested the same..

[Former Member]

Hi all,

siddhesh tawate,

I wanted to specify that the SLDCHECK is ok, i.e I can see the report and he's seems to be good.

I see all business system, integration server etc...

Raju Kurukuri,

I also check the rfc destination INTEGRATION_DIRECTORY_HMI, and the connection test is ok.

I have "error 500 internal server" in test result and in response Body, in details i have a series of number.

Guru,

I try to implement this note 1394710, but without success.

In fact, In this note I implement the pre-implementation manual abap step (step 1-7).

But I have the same issue in sxi_cache.

In this note I dont need to implement support pakage java since we are in 7.1 EHP1 SP3.

An idea ??

thanks by advance

[Former Member]

Once you implement the manual steps...you have to implement the note by using Tcode SNOTE ( which is not mentioned in the Note).. Then you will have success...

Let me know if you need further help

[Former Member]

ok thanks for information guru..

I will try this night at home in remote because I m in other client for the moment.

Regards

[Former Member]

Hi Guru...

I try to implement the note after manual pre implementation step, but i have a short dump "time exceeded"

I increase this parameter rdisp/max_wprun_time to 3600 but I still have a time out..

Best regards

N

[Former Member]

Hi Nor

Somhow we didnt face any short dumps... Is the short dump middle of implementing SNOTE or when you started Tcode SNOTE..if its the start, please check with your basis ( apart from setting the time limit) for any other reasons for time out...

If its in middle implementing SNOTE, Please re-check the 7 pre-implementations manual steps... Once they are good..try again SNOTE..

[Former Member]

Did you check RFC destination "INTEGRATION_DIRECTORY_HMI" in SM59? Just do a connection test and see if everything is ok.

[abhijitbolakhe]

Hi

Check wheather user PIAPPLUSER is locked?

unlock the service user PIAPPLUSER....

Regards

Abhijit

[Former Member]

hi, and thank you for all reply, but the issue wasn't solved and I already check all point of configuration which are describe here.

I check SM21 and default trace and i have few errors about the connection in my msserver who are KO.

You can see errors:

After having launched sxi_cache, i can see this log in SM21

SM21

An error occurred when attempting to update or refresh the XI

configuration data.

The error may have occurred when reading the business system, in the

communication layer, when retrieving the cache data, or when saving the

data. For further details about the error, call transaction SXI_CACHE.

Error Category: XI_CACHE

Error ID: BUSINESS _SYSTEM

Unable to determine the business system of the Integration Server. Test

the availability of the System Landscape Directory (SLD) by calling

transaction SLDCHECK or search for notes for the function module

LCR_GET_OWN_BUSINESS_SYSTEM.

Error ID: COMMUNICATION

Unable to establish communication with the Integration Directory. Check

the connection AI_JCO_DIRECTORY_SERVER by calling transaction SM59.

Error ID: OBJECTACCESS

Error when reading the Integration Directory data. Search for notes with

the subject Object Access in component BC-XI.

Error ID: UPDATE

The error occurred when writing the configuration data to the ABAP

cache. Search for notes with the subject ABAP Cache in component

BC-XI.

But the SLD chek is ok and Rfc connection too (in SM59 and NWA)

No user are locked and i already launched cpa cache, many restart, RZ70, check exchange profile, delete and create BS Integraion server but without success

Regards Nor

[Former Member]

Hi Nor,

Please check below note this should be helpful to you.

1104341.

When you say that SLDCHECK is successful is it because you see everything in green. Please note when you run SLDCHECK you need to see if the report displays the business systems correctly.

Hope this helps.

Regards,

Siddhesh S.Tawate

[Former Member]

HI Goudjil,

Please check in SLD, what is the last modified date for all XI/PI specific entries?

If PI/XI in sync with SLD then you have to see modified date as latest.

If not to all, as minimum as 4 entries.

1. Adapter Engine

2. Integration Directory

3. Integration Repository

4. RWB

If they are not of latest date then you can check below items before you do complete cache refresh

1. Check SLDAPICUST test works fine?

2. do once again RZ70 and then check if you have PI entries in SLD to latest date?

3. then if you are allowed perform complete CPA Cache refresh using below URL

http://<hostname>:<HTTP Port>/CPACache/refresh?mode=full

I hope these instruction would help in resolving the issue.

Regards

Sekhar

[Former Member]

My view ..just restart the server at this condition ...sure , will solve this issue .....

Regards

abhishek

Edited by: Abhishek Agrahari on Nov 4, 2009 10:08 AM

[Former Member]

check logs in sm21 , see the default log/default trace and figure it out. otherwise give restart.