[Roger Moss]

I was able to go back and spend some quality time on this problem and it turns out it was an issue with the backslash usage with cURL. I had to be specific with the pathing and use the backslash. 5.3 versions and before did not require this and the pathing was defaulting or getting set from the remote login if this was all you needed. This not longer appears to occur.

On a side note with our upgrade to 5.5Rev1 -AIX 5.2. We are very pleased so far with this release. We have our CPUs back! We were on 5.3 and the Hostserver was holding our Box hostage. We could only run just a few clients concurrently without it having a huge impact on cycles and memory. Now we see none of that. We are able to run many now with little to no impact. The SMAT load times are much improved also. We did see a change in the Concat function within the Xlate tool that we had to adjust to. So far that has been it.   Kudos to Quovadx for what appears to be a much improved release.

[Roger Moss]

I did this same interface last year and remember a few of the odd things about it:

– Dictaphone does not send an EVN segment on the transaction and it is required by HL7, so I had to make it optional on a variant.

– The rtf content in OBX 5 can and will exceed the default size, so I increased the size on my variant to 100,000 and had Dictaphone set a limit on their size to below that, which they can do. Also the field can repeat which is no big deal except that my receiving system could not handle that so in my translate I just keep appending it to a tmp variable with a CONCAT statement and then copy it after I fall out of the interation.

– And the last thing is that the Dictaphone side has a MAX message size that it can send on a message, so very large documents will get split across messages and I had to put a tcl proc on the TPS Inbound of the thread to write them out to disk and then put it all back together on a single message to the engine when the last one arrived.

Roger   8)

[Author]

Replies