[David Dillard]

Hi

I can give you some notes on what I have experienced.

We have integration setup between Phillips monitors and Epic, last I checked it was about 4 million msg per month.  We don’t save any SMAT for the reasons you mentioned, and have no message transformation going on, it is just a straight raw route of message between Phillips and Epic.
Messages into and across Cloverleaf flow smoothly, but we did run into issues on the Epic side sending all of the messages into a single AIP.  For some reason, for 5-10 minutes each day, there would be a delay getting the ack back from Epic.  This caused the outbound queue to skyrocket and take up to an hour to catch up, so they were seeing massive delays in the device data show up in Epic. I think it was some network issue, not Epic itself, but there was never a clear root cause for the delay so to overcome this we made some tweaks on the interface side of things:
The inbound processing was working fine, so we decided to create multiple outbound threads to spread the workload to Epic and do round-robin routing to send a message to each in turn.  This eliminated any outbound queues regardless of the inbound message volumes.
Multiple AIP’s were created in Epic  (copies of the original ) and new outbound threads added to each.
On the inbound thread:
Set the Trx ID Determination to be a new UPOC
Set up multiple named routes 0,1,2, etc
The UpOC keeps global count of which route to use (0 or 1 or 2 , etc)  and roll over back to 0 when the max number of routes is hit.

We set this up about 3 years ago and have not had any issue since.

Hope this helps

[David Dillard]

Did you try turning on the Disk-Based Queuing for the process?

We had a problem with large PDF messages causing memory issues but they seemed to go away after we turned this on.

The AIX 32-bit limit is a pain

[David Dillard]

I would create a translate to iterate over the OBR segments, save the provider information for each OBR then loop over all the OBX segment under the OBR and set the fields using the variables.

iterate over the OBR segments

     save the provider information into variables

     iterate over the OBX segments for each OBR

           Set the OBX fields with the variables

[David Dillard]

It is sometimes hard to get those details from Mckessson.

I have an example from the STAR v0.5 specs, but they have been known to change.

Hope it helps some.

ZOM – ORC/OBR extensions
Seq   Len   Data type   R/O/C   RP/#   Element name
1      4       SI         R        Set ID
2     15       CM         O        Alternate placer group number
2.1   15       ST         O        Alternate placer group number
2.2                      NS        Application ID
3      6       ST        NS        External lab location
4     10       ST        NS        External lab chart number
5     10       ST        NS        Lab order accession number
6      2       ID        NS        Lab administration site
7      2       ID        NS        Lab clinical processing site
8     75       CM        NS        Reflexor order[code]ZOM – ORC/OBR extensions
Seq   Len   Data type   R/O/C   RP/#   Element name
1      4       SI         R        Set ID
2     15       CM         O        Alternate placer group number
2.1   15       ST         O        Alternate placer group number
2.2                      NS        Application ID
3      6       ST        NS        External lab location
4     10       ST        NS        External lab chart number
5     10       ST        NS        Lab order accession number
6      2       ID        NS        Lab administration site
7      2       ID        NS        Lab clinical processing site
8     75       CM        NS        Reflexor order

[David Dillard]

I wrote this script to check the processes in our site and send out an e-mail if any of them were dead or down.  A cron job was used to run the script on a regular schedule (every 15 or 30 min depending depending on the time of day)  

As David pointed out you want to avoid “crying wolf”  — the e-mails either get everyone in a panic or are ignored as false alarms.

You should be able to modify it for your specific site and e-mail address and adjust the text to read how you like.

[David Dillard]

Sorry for the confusion

— I did not read your question close enough and my mind took off on it’s own tangent.

I plead caffeine deficiency

[David Dillard]

Hi,

You can also substitute a CR-LF for every CR in the message using the regsub command

Code:

regsub -all “r” $msg “rn” msg

Like David said, run this on the inbound data before any other formatting or translations take place.

Here is a sample proc to get you started

Code:

proc Add_LF { args } {
   keylget args MODE mode                  ;# Fetch mode
   set dispList {}             ;# Nothing to return

   switch -exact — $mode {
       start {
           # Perform special init functions
       # N.B.: there may or may not be a MSGID key in args
       }

       run {
           # ‘run’ mode always has a MSGID; fetch and process it
           keylget args MSGID mh
           set msg [msgget $mh]
           # substitute all with
           regsub -all “r” $msg “rn” msg
           msgset $mh $msg
           lappend dispList “CONTINUE $mh”
       }

       time {
           # Timer-based processing
       # N.B.: there may or may not be a MSGID key in args
       }
       
       shutdown {
       # Doing some clean-up work
   }
   }

   return $dispList
}

[David Dillard]

HI,

Here is a script I used to e-mail the status of the processes and connections on the engine.  It only sends mail if a process or connection is dead or down.

I had this run as an hourly report from a cron job.

It should give you a place to start:

Code:

#!/bin/ksh
###########################################################################################
# File:  mailenginestatus.sh
#
# Author:  David D Dillard
#
# Description:  This script is used check the process and connection status of the
#               QDX engine and send a mail to the notification group with the status
#               information
#        18jan2005-ddd modified to only send mail if nothing is dead or down
#
###########################################################################################
site=”live”
sitepath=”/quovadx/integrator/liveroot”
# If script is not running in the site this will swap into the correct site
#   Edit to add whatever scripts you use to swap to your site
if [[ $HCISITE != $site ]]; then
       setroot $sitepath $site >> mailenginestatus.log 2>&1
       cdsite
fi

#now that we are in the correct site we can proceed
mailto=”enginesupport@company.com”
subject=”QDX alert message: Status report for the $site interface engine site”
echo “QDX engine status report for the $site site” > mailenginestatus.log
date >> mailenginestatus.log
echo “nn————————————————————————-” >> mailenginestatus.log
echo “The following processes are dead” >> mailenginestatus.log
hciprocstatus | grep dead >> mailenginestatus.log
echo “nThe following connections are down” >> mailenginestatus.log
hciconnstatus | grep down >> mailenginestatus.log
echo “nn————————————————————————-” >> mailenginestatus.log
echo “Complete hciprocstatus report:” >> mailenginestatus.log
hciprocstatus >> mailenginestatus.log
echo “nn————————————————————————-” >> mailenginestatus.log
echo “Complete hciconnstatus report:” >> mailenginestatus.log
hciconnstatus >> mailenginestatus.log

tmp=”/tmp/livemailenginestatus.tmp.$RANDOM”            ;# Create random temp file
hciprocstatus | grep dead >> $tmp 2>&1
hciconnstatus | grep down >> $tmp 2>&1

# search the status for any dead or down processes or connections
if (grep -c “d” $tmp > /dev/null);
then
    mail -s “$subject” $mailto < mailenginestatus.log
fi
sleep 2
rm mailenginestatus.log
rm $tmp

[David Dillard]

We regularly use purge cache and have not noticed any problems

Running: Version 5.3 rev2

David

[David Dillard]

I had a problem that sounds similar to what you are experiencing,   I would double click on an item in my xlate but it appeared nothing would happen.

In my case the properties window was popping up but it was off my desktop — when I switch to a higher resolution I could see the window on the far right of my screen.

I am using version 5.3 so it may vary but here is one way to fix this:

Look in the folder C:quovadxqdx5.3integratorclient (or wherever you installed quovadx)

There is a file named client.ini

Open this up in notepad and look for a section with your username in it similar to this:

[user_yourusername_xlate_config]

On my computer it looks like this:

Code:

[user_davdil_xlate_config]
property_dlg_pos=48 197
property_dlg_size=827 409

The property_dlg_pos controls the screen position (with coordinate 0,0 in the upper left of the screen).  The property_dlg_size controls the size of the property dialog window.

If the property_dlg_pos coordinates are very large, or even negative the window is probably off the visible screen.

Change the values for the property_dlg_pos to move the window to where you can see it.

I would backup the client.ini file before making any changes just in case something goes wrong

This may not have anything to do with the problem you are seeing but thought I would toss it out there just in case

[David Dillard]

Hi,

We send mail out from both alerts and tcl procs using the unix mail command.

From an alert, use the exec action.  Echo out a mail message and then pipe it to the mail command:

Code:

exec {echo “this would be the body of the message” | mail -s “%A”  address1@mail.com,address2@mail.com }

The %A variable contains information about the alert that triggered.

---

From a tcl proc you can do the same thing:

I usually setup arguments that will let us pass the email arguments into the procedure so we can change them without making changes to the code:

Code:

set email1 “default@mail.com”
keylget args ARGS.EMAIL1 email1
exec echo “body message” | mail -s “Subject” $email1

More than one email address?

Code:

exec echo “body message” | mail -s “Subject” “$email1 $email2 $emaill3”

[David Dillard]

I just needed look in the right direction.

Once I took a look at the inbound TPS we currently use to ack the client it all became clear.

Thanks everyone for the guidance

[David Dillard]

Ryan,

Thanks for the ideas.   Something must have become ‘hosed’ in the thread’s pending messages, but when you mentioned “resent” messages it gave the idea to save off any messages in the recovery database and then clean them out to start fresh.

It did the trick and now things are working fine.

Thanks for all your help

<>

---

Just for future reference (my own mostly) , I did do some additional troubleshooting

The outbound save message really had x0a after each segment.

Code:

000838f0 7c525553 5045440a 4f42587c 347c5458 |RUSPED.OBX|4|TX
00083900 7c363832 3134345e 4e2e474f 4e4f5252 |682144^N.GONORR
00083910 484f4541 45204259 20544d41 7c317c4e HOEAE BY TMA|1|N
00083920 45474154 4956457c 7c4e6567 61746976 EGATIVE||Negativ
00083930 657c7c7c 7c467c7c 7c323030 35313032 e||||F|||2005102
00083940 35303834 337c4f4d 4c5e5e5e 30353130 50843|OML^^^0510
00083950 32343036 3334367c 52555350 45440a4e 2406346|RUSPED.N
00083960 54457c31 7c7c4e65 69737365 72696120 TE|1||Neisseria
00083970 676f6e6f 7272686f 65616520 62792054 gonorrhoeae by T
00083980 4d412069 73204644 41206170 70726f76 MA is FDA approv
00083990 6564206f 6e6c7920 666f7220 656e646f ed only for endo
000839a0 63657276 6963616c 0a4e5445 7c327c7c cervical.NTE|2||

At one point I thought it might be the the delivery protocol so I made a copy of the thread and routes, keeping everything except setting the copy to deliver the message to a file

The message in the copy was correct, so I could eliminate the xlate, tcl procs and other settings in for the route and thread.

I knew the protocol was working for other threads without a problem so that led me to look a the actual data.

The data format was ok for the thread copy – so that (along with Ryan’s suggestion about resent messages) let me to consider something in the pending message was causing the trouble.

[David Dillard]

Hi,

We also ran into some problems with a couple of our connecitons where they get ‘clogged up’ and stop sending messages.  The root cause could not be located so a ‘temporary but semi-permanent fix’ was put in place.

An alert was setup to watch the thread and to call an external script when triggered.

Example:

{VALUE opque} {SOURCE Test1} {MODE actual} {WITH -1} {COMP {> 10}} {FOR {nmin 30}} {WINDOW {* * * * * *}} {HOST {}} {ACTION {{exec {bumpThread.sh Testoutbound Test1}}}}

The script then uses the hcicmd to stop and start the thread.

We created a generic script to allow us to pass in the process and thread names into the script so it can be reused

$1 == Process Name

$2 == Thread Name

#; stop the thread

hcicmd -p $1 -c “$2 pstop”

#; Set for specific timespan but could be changed

#;     to watch and wait for the thread to go down before

#;     bringing it back up again

sleep 30

#; start the thread

hcicmd -p $1 -c “$2 pstart”

[David Dillard]

HI,

We do an order consolidation process on some of our inbound order messages which sounds similar to what you are after.

To do this we use two threads:

[Author]

Replies