Re: serial IOC connection problem

[Mark Rivers via Tech-talk <tech-talk at aps.anl.gov>]

Hi Oksana,


It may be that this device wants to just do a single transaction per connection.  That is how Web servers work.


drvAsynIPPort supports that mode of operation directly.  https://epics-modules.github.io/master/asyn/R4-41/asynDriver.html#drvAsynIPPort


You need to change this line:


drvAsynIPPortConfigure("FTS", "10.16.2.195:49151")


to this:


drvAsynIPPortConfigure("FTS", "10.16.2.195:49151 HTTP")


Once you do that I think you can remove the "disconnect" commands from the protocol file.


If you still have trouble please turn on ASYN_TRACEIO_DRIVER for the IP port and send the output.


Mark





________________________________
From: Tech-talk <tech-talk-bounces at aps.anl.gov> on behalf of Ivashkevych, Oksana via Tech-talk <tech-talk at aps.anl.gov>
Sent: Wednesday, December 9, 2020 12:51 PM
To: tech-talk at aps.anl.gov
Subject: serial IOC connection problem


Hello,

I have been asked to port an existing Serial IOC  for a Force Torque Sensor from ATI from one server (Debian 7 with Debian packages dated ~ at least 5 years ago) to CentOS 7.



CentOS 7.7.1908; base R7.0.3.1; asyn R4-37; SerialDevice master.



After the ported recompiled IOC reported unrecoverable connection error, I have created a brand new IOC using the latest makeBaseApp.pl script.



Errors became more dynamic, and I do see some readings, when IOC does manage to connect successfully. I have browsed the Tech Talk and found a related discussions about connecting/disconnecting. Protocol file of this IOC does  connects/disconnects after every read and is very simple. I only can guess why the developer made this decision, I am sure it was an unavoidable compromise.  I did add asynSetTraceMask call which changes temporarily the dynamic of the errors.  Logs and proto are attached.



In summary: I have two iocs accessing same device: the old one running on Debian 7, and the new on running CentOS 7. The old one works without problem the new on reports connection errors and does manage to read something at the frequency ~30 times slower than 0.1 expected scan. I am out of my own ideas, and would appreciate help from the community.



Thank you,

Oksana