on 10-03-2012 12:50 PM
Hi,
I created an integration process in our development environment and it worked perectly. However, for reasons unclear one day it just stopped working and all requests to it timed out. I therefore simply copied the process and created a new series of configuration objects to match. Again it worked perfectly. This shows me there is nothing wrong with the IP. After some thinking I suspected what might be happening is:
IP works fine
Message is passed into IP which causes an error
Error generated is unhandled
IP stops working
i.e. if the IP gets an unhandled error it appears to get stuck in some sort of error state and is unusable.
To prove this I took a working copy, passed invalid data to it and indeed it did break it.
I will need to look at ensuring there are no unhandled errors in future. In the meantime my question is....how do you get the IP out of this apparent error state ? Is there something you can clear. At the moment when an IP stops working due to an unhandled error it never works again. This can't be right. There must be something I can do to get it working again.
Cheers,
Paul Clements.
Hello Paul,
have you analyzed the error for the first message that failed? I have made the observation that all messages for the same scenario get queued up in the same queue for the IP. So when the first one fails due to wrong data, some mapping error, whatever, all subsequent messages would also fail. If this is the case, generating a parallel scenario with the same objects under a different name would of course work as before - until the next message hits an error.
If this is true you could see a blocked queue in SMQ2 that looks like an EOIO queue.
Regards,
Jörg
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Paul,
you noticed a well know critical behaviour of ccBPM. Therefore it is actually mandatory - not from a technical point of view, but from a business point of view - to define exceptions and the corresponding exception branches for each transformation and send step.
The "error status" process after a not handled exception is annoying most seriously during use of correlations, because the hanging process still correlates. You can stop that nonsense by deleting the work items in SWWL.
Regards,
Udo
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
1. make sure you are using excepton handling for all possible failures within the IP. this is significant to maintain a healthy IP
2. Go to SWWL tcode and delete the error WP and the try the scenario again
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
84 | |
24 | |
12 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.