![]() |
![]() ![]()
Experimental Physics and
| ||||||||||||||||
|
we have seen the same problems with the V2.0 Gateway running on Linux. I was able to cut it down to a small Linux system with just one analog record running on a soft IOC and a Gateway running on the same machine exporting this channel on a different port number. Starting camonitor on a different host to directly watch the channel on the IOC I was able to see random junk values being written into the channel as soon as two clients are writing fast (slider) through the Gateway. Not just jumps (that I'd expect), but really random bit patterns that are interpreted as doubles. No matter which system the clients run on (tested with HP and Linux). Connecting the clients directly to the soft IOC - no problem. At BESSY it didn't kill hardware as our DRVHs are set properly ... I hope. We lost the beam a couple of times until - as an instant measure - we stopped running write accesses through the Gateway. Read accesses seem to work fine. The V1 Gateway does not show this behaviour - but we had to switch to V2 because the old hardware died. I was already reporting this problem some months ago, but the APS installation hardly uses write access through the Gateway (if at all), so Ken doesn't see these problems - which makes it hard finding the bugs as well as the motivation to fix them. But again: One Linux box running one ao channel on a soft IOC, the Gateway, the camonitor, and two simple clients is enough to reproduce the problem. (On my RH8 Linux box, at least.) Hope this helps (well, I know it doesn't really, but ...;-) Ralph Rolf Keitel wrote: **** Has anybody else experienced these problems? ****
| ||||||||||||||||
ANJ, 02 Sep 2010 |
![]() · Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing · |