Hi all,
Thanks for these replies. It is very possible we have newer Cisco switches installed
at various places throughout both our telescopes. I will try to find out the models and
see if we can capture packets on this one troublesome instrument (EPICS 3.14.8.2).
This thread has evolved. So I want to be clear that there may be two similar issues here:
1) In the original post I reported a problem how a script on linux calling caget once per
second can timeout talking to our new Telescope Control System (TCS) IOC running
RTEMS-4.10.2/EPICS-3.14.12.7. We couldn't find such timeouts on the legacy
TCS running VxWorks 5.4/ EPICS-3.13.9. When these timeouts occur, we _DO NOT_
see any log messages on our TCS. I'll try to use Michael's cashark plugin to learn more.
2) Then, my colleague Ignacio at Gemini South posted a similar case where one of
our instruments (3.14.8.2) calls caget and often fails with the new TCS. Every time this caget
fails we see this:
Jul 12 12:40:59 E) PORT: tcs_vme, MSG: CAS: Request from 172.17.102.130:57113 => cmmd=12 cid=0xffffffff type=0 count=0 postsize=0 Jul 12 12:40:59 E) PORT: tcs_vme, MSG: CAS: Request from 172.17.102.130:57113 => available=0x1 N=0 paddr=0x0
There's a Holiday in Chile today but I hope they can arrange to capture some packets
in the next day or two. Details are forthcoming.
Thank you again,
-Matt