One theory is that the alarm server is overloaded and can't process the tag subscriptions in time (alarm scan time defaults to 500ms). Or it could be on the IO Side, it isn't providing the data in time, e.g. it is stale data being provided to alarm server. Check the CPU and memory usage of Alarm server IO Server. Try adjust the parameter [Alarm]Scantime to say 1000ms and see if it makes any difference.
Olivier Thanks for the suggestion, I think this issue is being caused by an overloaded IO server (or IO device). Other symptoms occur when viewing a page with a lot of tags, the objects on the page are constantly flashing #COM. I tried setting [Alarm]Scantime to 5000 but this did not resolve the issue. Shutting down the alarm server will resolve the flashing #COM issue on the page however.
I thought one solution would be to reduce the page scan time on the busy page, however in all my tests I haven't seen any change in the update rate of graphics on the page, as if the setting is having no effect. I've tried:
The last two methods did at least result in the reported scan time (returned from PageGetInt(-2) run from the context of the problem page, via kernel) matching the set scan time. This was not the case when I set it via the page properties. In all cases though the genie animations were updating much more frequently than every 5 seconds, so I'm doubting the effect of this parameter.
Hi Philip,
In a project we were on a couple of years ago, support indicated that the [page]scantime needs to be changed in conjunction with 2 other parameters in order to achieve the desired effect:
[page]anmdelay and [animator]flashtime
Hope that helps
Stuart
Thanks Stuart, that did seem to slightly reduce the frequency of the animation. I also confirmed in csatopsi.subs that the majority of tags were now subscribed with an update rate of 5000ms. The issue still remains, but at least I I know more about page scan times!