EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: IOC and its OS health monitoring
From: Dohn Arms <[email protected]>
To: <[email protected]>
Date: Thu, 22 Dec 2016 14:05:31 -0600

Recsync is now being maintained by Dylan Maxwell.

https://github.com/ChannelFinder/recsync

Recsync is intended to maintain a authoritative central database of all
records provided by all IOCs.  In order to be authoritative, each IOC
maintains a persistent TCP connection to a central server.  This gives
some feature overlap with aliveRecord*.

Zero configuration is a goal.  No per-IOC config is required.  Just link
in the reccaster support module.  There is an optional .db which
provides connection status.

While intended to be used with channelfinder, it is not tightly coupled,
and could be used with other databases.


* In my ideal super efficient world, aliveRecord would use the recsync
protocol.


The alive daemon I maintain for my group has zero configuration as well. Just point the alive record to the daemon, and it adds the IOC to the database, and infrequently probes for information. All the information for doing this is in the alive packet.

UDP packets are used because if the packets get lost, the IOC doesn't care. There's no worrying about reconnecting if the server goes down, as there is no persistent connection with UDP. The design of alive is to minimize the affect of the remote server on the IOC. The only direct connection allowed from the server is in the form of allowing a connection on a TCP port, alive shoves its information into it, and immediately closes it. The server can't write to the IOC or even keep a port open.

With UDP, I have all UDP packets incoming onto the same port on the server, and the same thread just keeps reading packets off the same port. If the IOC information needs to be read using the TCP port on the IOC, I just spawn a temporary thread for that purpose. There is also no buffer allocated for UDP. With TCP, each connection get passed off to a new port, needing a separate socket which also has its own buffer attached to it. As we have hundreds of IOCs talking to the server, I don't want a TCP connection.

So, alive is not switching to TCP....

Dohn


References:
IOC and its OS health monitoring Shen, Guobao
Re: IOC and its OS health monitoring Shen, Guobao
Re: IOC and its OS health monitoring Michael Davidsaver

Navigate by Date:
Prev: Re: IOC and its OS health monitoring Michael Davidsaver
Next: Re: Alive record server Dohn Arms
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: IOC and its OS health monitoring Michael Davidsaver
Next: Acromag IP230A/231 support Ron Sluiter
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 22 Dec 2016 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·