Problem with ‘Await Replies’

[garry r fisher]

Hi,

I’ve logged this call with Quovadx as well but I thought I’d seek a broader view hence this.

I have Windows 2K running QDX v5.2 – a single process in a site running one uni-directional HL7 interface. Await replies is set and the timeout is currently 30 seconds (it was 5 but we changed it to prove something – reado on). We have hcitpsmsgkill set in the inbound replies and are using the standard mlp pdl.

When we send a message and receive the MSA ACK back the next message is sent only when the timeout is reached. Originally it was 5 seconds which could have been coincidence so we changed it to 30 seconds and it is sending a message every 30 seconds even though the ACK is been returned in seconds.

I have run a full engine output and everything looks OK except for each message waiting for the timeout.

Has anybody seen anything similar or have any suggestions as to how I can resolve this.

Thanks

Garry

Topic

[Daniel Lee]

Is the hcitpsmsgkill the only proc your using in the tps inbound reply or are you using the ValidateReply and the KillOBSave as well?  I could make sense out of it if you were validating your replies but can’t think of an answer if your not.

[Richard Hart]

Garry.

When you are killing a reply message, you need to use the disposition ‘KILLREPLY’ rather than KILL.

The hcitpsmsgkill performs this for you, as long as the reply context is sms_ib_reply.

eg from the scipt …

if [cequal $context sms_ib_reply] {

   set disp KILLREPLY

} else {

   set disp KILL

}

Are you sure you are killing the mesage from the correctmessage context?

[garry r fisher]

Daniel/Richard,

I’m only using hcitpsmsgkill although one option I have is to use the recovery_38 procedures to see if that makes any difference – unfortunately we need to clear the backlog first and that’s currently 17000+ messages.

I used ‘enable_all’ eo config to check the inbound replies but they were being killed before they got to the engine (obviously) so I’m not sure what context is being detected. I might mod the tps to echo a status to check.

Any other thoughts while I do these?

Garry

[Daniel Lee]

Are you saying that when you used “enable all” you couldn’t see the ACK coming back to the thread in the process log?

[garry r fisher]

Hi,

Sorry- I can see the ACK been returned to the thread but as it is been KILLed it is not been ‘dumped’ in the long format.

G

[Author]

Replies