Re: CA client broadcast brings down galil which is not running EPICS

[Henrique Almeida <henrique.almeida@lnls.br>]

 I also forgot mentioning that channel access ports are 5064 and 5065. There's no IK command parameter you could use to block it.

2015-05-26 14:04 GMT-03:00 Henrique Almeida <henrique.almeida@lnls.br>:

 Given your explanation, it looks like you're not protecting the galil controller from the network broadcasts. Galil listens to TCP ports 0, 23, 68, 502 and from 1000 to 65535 and UDP ports 0, 23, 68, 502 and from 1000 to 65535. If the command IK is set to 0, all the TCP and UDP ports 0 to 65535 are open. Any broadcast on the network will be interpreted as a command. Since galil firmware and protocol are not robust, there are invalid commands that will make it halt, only recovering with a restart.

 You should use galil only on it's own, separated subnet, or you should add rules to your switch to block all ports targetting galil, except the one you use. Or you can give up ethernet and use USB, serial, etc.

2015-05-24 0:48 GMT-03:00 Mark Clift <Mark.Clift@synchrotron.org.au>:

Dear All,

Firmware revision C is now available for the 40x0 controller, and the release notes include several ethernet fixes. It is worth trying this firmware C. All clues regarding cause are important (CA broadcast etc).

We could ask Galil to make the IK command block CA ports? If EPICS CA broadcasts are what is causing the Galil to become unresponsive on the network, enhancing the IK command would seem the way to go.

Have you tried a simple IK1 command on the controller? Please refer to command manual for IK command. This may help, but others have reported it didnt help at all. With an older RIO 47200 this fixed our problem with it dropping off the network.

A cause needs to be found, before we can describe a fix.  Galil have known about this problem for some time, but are unable to reproduce it in their lab. Galil are keen to find a fix, but may need help as their test environment does not include EPICS.

JinHu, I'm most curious to hear your test results later in the week. I hope you can find the cause for us.

Best wishes,

Mark

Dr. Mark Clift

Senior Controls Engineer

Australian Synchrotron

800 Blackburn Road

Clayton 3168

Ph: +613 8540 4264

Fax: +613 8540 4200

Mail: Mark.Clift@Synchrotron.org.au