Handling Exceptions When End System are down

Hi All,

Thanks for checking my Thread…

I’m working working on a flow which synch data from oracle db and SAP R/3 sys… its a plain intergration , No Monitor here…

  1. How can i handle the exception fired during the end systems are down… I mean when wM is holding data pulling from db if SAP is down … what i can do… In SAP i’m invoking an RFC wic will push data into R/3 sys.
  2. Similarly in vice versa when i need to push into db … what can i do…
Solns:

As i know i can have Retry option … but if it fails in all retries also then?

And how actually configure Retry option in flow service…?

Help is greatly appreciated :slight_smile:

Thanks…

Rgds
Shaik

Comments are as follows :

  1. Have the DB operation and SAP RFC call as a single block of operation . When the end systems are down the worst case is half data-synch’d and half left over. Address this issue first .

  2. Use batch insert/select for DB .

  3. In both the cases (to and from DB ) you should be using canonicals and those canonicals can be serialized to your disk in case of any failures and you use a retry service to manually start from the point where it failed .

  4. Retry can be done by using repeat . Exit on success , retry on failure . You many dynamically give the values for retry count and retry interval as per your need .

Thanks

Thanks Dhruv,

I’m using Pub-Sub Model wit Guaranteed Delivery… and What the use of Batch insert here?
There are no Canonical used here …

This Retry option i know…

I wanna know there is a mechanism … In case of Transient error how can i handle… there is some option … to throw a retry exception checking the flag…

Rgds
Shaik

This has been covered several times in this forum. Do some searching and you will find the answer. It’s also covered in the pub/sub guide that comes with the product.

I have searched all the posts … but none was suited for my prb; And If wM Monitor is used then it easy … but there is no monitor here…

Regards
Shaik

Have you considered using pub/sub?

Retries are configured on the triggers. Just make sure your triggered service throw a ‘Retry for Exception’, and that they are transactional. You can have infinite retries - of course, your remote system would have to come up before the queues filled up.

I’m given to understand that you really need Broker to make this enterprise grade.

Read pub/sub documentation , and look for resource monitoring service . U need to create a resource monitoring service and configure your trigger accordingly.read the documentation so that u can understand clearly.
HTH
sri