Message stuck in DLNG status in Mail Receiver

[Former Member]

Hi All,

We are facing an issue where the first message sent after a long time (or the first message of the day) gets stuck in the DLNG status.

However, the messages sent after that get delivered promptly.

Can you please help with figuring out what the issue is and how to overcome this?

We are using SAP PI 7.31 Single Stack.

Thanks in advance!!

[Former Member]

Hi All,

I have changed Sync-Async bridge scenario to normal Asynchronous scenario.

Thanks,

Mohammed Gouse

[Former Member]

Hi Mohammed,

I think, the issue might be with SMTP services availability and mail traffic. Check with basis team for SMTP service availability during that time.

Regards,

Krupa

[Former Member]

Hi Krupa,

Thanks for your reply!

The very next message that was triggered just a few seconds later got delivered.. So I doubt that it has anything to do with the SMTP services.

Regards,

Gouse

[Former Member]

Hi Mohammed,

Make sure that the SMTP services up and running.

Regards,

Krupa

[Former Member]

Hi Krupa,

I am Gouse's team mate.Yes, the SMTP services are up and running. The Messages triggered after the first message have all been delivered. However, the first message is still in DLNG state (it has been over 3 hours now!)

Regards

Vikram

[engswee]

Hi Vikram

The worker thread used for that message might have hanged. Try and stop the corresponding receiver channel to release the worker thread, and then resend the messages.

Rgds

Eng Swee

[Former Member]

Hi Vikram,

Also, check the payload for To email id and size etc...

Regards,

Krupa 

[Former Member]

Everything is exactly the same for the messages (4 messages as seen in screen-shot)

Regards,

V

[Former Member]

Hi Vikam

Messages goes to delivering state if the thread that is executing the message get hanged. This can happen due to system restart, server down for temporarily while the message is in still processing state.

If it's only a single message, then cancel it and try to resend again in PI. If there are multiple messages then

1. Try to restart the mail adapter service in NWA

or

2. Restart the java stack which is the last option.

[Former Member]

Hi Eng Swee,

Thanks for your reply!

None of the threads are hanging - which is what is so strange!

This is happening only for the first message of the day and not the subsequent messages!

Really struggling to understand this behaviour!

Regards,

V

[Former Member]

Hi Indrajit,

Thanks for your reply

It is only one message - and that too the very first message for the day (we observed a similar behaviour yesterday, too)

When we send messages seconds after the first one, it gets delivered without any issue!

Also, the message eventually gets delivered after a long time (this happened once before - but today no luck yet)

[Former Member]

Hi Eng Swee,

I was wrong about my previous reply.

I was checking only the mail adapter threads - where none of the threads were being used:

But, when I scrolled down to SOAP, i see that there is one thread that is stuck since yesterday and no message in the queue

[engswee]

Hi Vikram

Did you check the threads over all the different server nodes available in your PI system? There are at least 2 server nodes on a typical PI installation.

I don't think the SOAP worker thread is related to the hanged message in the mail adapter. It could just be another thread that was not released properly.

Rgds

Eng Swee

[Former Member]

Hi Eng Swee,

This is the Dev system - so only one server node.

And it looks like the SOAP thread hanging is the issue!

Ours is a Soap-to-Mail scenario and when I was browsing through this fantastic blog, I came to understand that in a Java only (ICO) scenario all the processing steps in the Messaging System are executed in the sender queue. The receiver queues are not used at all.

Right now, in the Dev system, we are no longer facing the issue - need to see if anything has been changed in the system since last we checked.

We were, however, facing the same issue in the QA system as well - need to check if the issue has been resolved there as well.

I'll post my finding soon!

Thanks for the replies - it helped us look a little more diligently!

Regards

V

[engswee]

Hi Vikram

Yes, you are right, for ICO the worker thread is used on the sender side (I sometimes forget about this as I've been so used to thinking in terms of classical scenario!!), so it totally makes sense for your scenario.

Hope things work well in your QA system. Please ask Gouse to close this thread if everything goes well.

Rgds

Eng Swee