Peculiar Cache problem

[Former Member]

Hi,

We are testing an interface involving 2 back to back synchronous RFC calls.

There is no issue with the interface except for sometimes when it throws up an error in SXMB_MONI about a problem in the message mapping . (Transformation error). But when we check it in the IR after copying it from the payload, there is no problem in the message mapping and moreover, the mapping doesnt fail all the time. We suspect that XI is reading older version of the cache.

When the same file is processed again , it goes through. Whenever , we clear the XI adapter meta data cache for that particular mapping, the interface works. Immediately ,if we process the same file again, it throws up the same error. But if we process the same file after a gap of , say 5 min, there is no problem.

This is a high priority issue that has been lurking around us for quite some time. We have already tried various options like SXI_CACHE refresh, CPA cache refresh, checking cache notification for errors etc etc.

What else could be the reason for this issue?

Any input in this regard could be really helpful at this stage.

regards

krishna

[Former Member · Accepted solution]

change mapping and try to activate it again and make sure that the cache is up to date..

Rajesh

[Former Member · Accepted solution]

Hi rajesh,

we have actrivated the objects already. but same issue.

regards,

krishna

[JoelTrinidade]

Hi Krishna,

Try setting up the refresh time of the cache. Refer:

Cache Refresh Errors - New Phenomena

/people/sravya.talanki2/blog/2005/11/03/cache-refresh-errors--new-phenomena

SXI_CACHE - Ripped Off

/people/sravya.talanki2/blog/2005/12/02/sxicache--ripped-off

Rgds

joel

[Former Member]

Hi joel

Where to set up the refresh time for the cache . Do u mean to enable the "No Timeout" setting in the HMMI_ IntergrationDirectory setting (sm59)

regards

krishna

[JoelTrinidade]

Yes krishna, please try out that option or the other two over there since you have this problem on immediate run and if you do not run immediately but wait for five mins then everything is fine. So its a issue with time out.

Rgds

joel

[former_member206760]

check for locked Adapter Entries in Visual Administartor

[Former Member]

@ TArang

No locked adapter entries

regards

krishna

[former_member181985]

Hi,

Try increasing the timeout parameter for RFC Adpater using Visual Admin. I think this not related with cache.

The mapping error comes since the Adapter framework returns a different response other than RFC response bcoz of timeout.

Thanks,

- Gujjeti.

[Former Member]

@ Praveen

We dont think the issue could be because of an RFC timeout since the whole interface takes just 5 secs to complete.

btw, We did not find any RFC time out parameter in Visual admin

[former_member181985]

what is the data size for request & Response?

Any tunnels in between the RFC sender & RFC receiver..........

[Former Member]

@ praveen

Data size <500 kbs

No tunnels

regards

krishna

[former_member181985]

Not sure if this helps.............

Check Michal's blog to increase RFC timeout

[/people/michal.krawczyk2/blog/2006/06/08/xi-timeouts-timeouts-timeouts|/people/michal.krawczyk2/blog/2006/06/08/xi-timeouts-timeouts-timeouts]

[Shabarish_Nair]

 This is a high priority issue that has been lurking around us for quite some time. We have already tried various options like SXI_CACHE refresh, CPA cache refresh, checking cache notification for errors etc etc.

definitely a cache issue. Try working on fixing it - https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/c0332b2a-eb97-2910-b6ba-dbe52a01...

as a current alternative, for the time being copy the mapping to a new name, refer the new one to the interface mapping and run tests.

[Former Member]

@ Sabs,

Thanks. We will do it but what could be the cache issue exactly ?

Krishna