Strange "no-match: no more phrases to try" issue

[Terry Kellum]

We did an upgrade of our lab system hardware, and while we were down, I turned off the threads outbound to our LIS, and inbound from our LIS during the downtime.  This allowed the 5300 inbound messages to stack up against the outbound thread (to the LIS).

Since coming back up with the LIS, we’ve been seeing these “no match” errors on the inbound engine thread on the connection from the hospital.

To troubleshoot, I named in Rob’s wonderful mlp_tcp-showNoMatch PDL.

Looking at the EO, I see message fragments, usually mid-message to the end of the message.  I can usually see the encapsulation chars at the end of the message fragment (0d 1c 0d) but don’t see the 0b at the beginning of he message.

Today, I’m looking closer and I see stuff like this…

Code:

(bunch of binary HL7 with various patient info)
7c 7c 7c 7c  7c 7c 7c 45  ||||||||E|
56 7c 5e 5e  32 41 44 44  |V|^^2ADD|
49 41 47 31  5e 5e 43 48  |IAG1^^CH|
45 53 54 20  50 41 49 4e  |EST PAIN|
0d 1c 0d 0b  4d 53 48 7c  |….MSH||
5e 7e 5c 40  7c 50 4d 53  |^~@|PMS|
7c 4d 42 33  7c 7c 7c 32  ||MB3|||2|
30 31 30 30  37 32 38 31  |01007281|
36 34 39 7c  7c 41 44 54  |649||ADT|
5e 41 32 38  7c 43 48 50  |^A28|CHP|
46 53 45 4e  4f 7c 7c 32  |FSENO||2|
2e 33 0d 5a  53 48 7c 52  |.3.ZSH|R|
41 44 7c 53  4f 41 52 49  |AD|SOARI|
41 4e 0d 45  56 4e 7c 41  |AN.EVN|A|
(The rest of the message with no terminator)

I’m thinking a possible PDL timeout issue, so I’ve modified the PDL and added a “timeout:60000” to the hci_pd-msg_style clause.

It’s baking right now.

Has anyone seen a PDL timeout issue, and do these symptoms fit?

I’ve done the usual network plug, reboot switch etc.

I also took the site down and rebuilt my database to get the tablesize down. (hcidbinit -A)

I’m running CL 5.6 on RedHat 5 on VMware 3.5

Topic

[Michael Hertel]

Look for a 1c without a trailing 0d.

I’ve seen issues where we’d receive datadata

The errant 1c in the data throws things off.

[Richard Hart]

I think that this is the error caused when correctly encapsulated messages are are sent to Cloverleaf without waiting for an ACK back from Cloverleaf.

[Charlie Bursell]

Not really true Richard.

Whatthe error means is the PDL has not been able to find a match for the dsignated pattern.  For example, the MLP driver looks for a pattern like:

    X0BX1CX0D

Turn u EO and look at the messages.  My bet is you will see one of these missing.  Another possibility is having either a Hex 0B or Hex 1C within the data

[Terry Kellum]

It turned out to be a bad fiber transceiver in the path between our cloverleaf engines.  We were running into timeout issues and dropped data.

Our data stream was basically fruit salad and Cloverleaf was not liking the buffet we were handing it!

We replaced he transceiver, and Cloverleaf started humming along, like it always does.  Performing it’s job and not giving us anything to complain about!

[Author]

Replies