I have a system running on 7.8.4 that is experiencing clients locking up ie needing a to alt+cntrl+delete out of the screen, and a gateway that is sometimes showing over 100% usage like this
It seems like the gateway restarts itself from CPU usage, though I haven’t personally seen it happen, the systems says the uptime was only for about 4 hours so it definitely restarted around 8 am and I don’t believe the computer itself was restarted or touched for any reason.
The gateway server has 12 gigs of total ram of which the .conf file says it can use 6 gigs and initalizes with 3 gigs, as well as 4 3.2 gHz processors.
The project has 7 transaction groups running, each of which has 5 tags being written to the database at relatively slow rates (30 seconds).
The project itself is just for data viewing, it has graphs and tags all over on different windows, but no real input from users (it’s only supposed to be used to display information on tv monitors around the plant or from office computers). I’ve only seen up to 4-5 clients in use at one time in my time here though I am told it can be as many as 20 (though I’ve seen the performance issues with only 4). There are times when the clients go from 4 to 0 back to 4 - I suspect this is due to gateway disconnects happening when the server is at CPU or RAM capacity but I don’t know for sure.
The project properties for clients is 64 MB initially and 1024 MB.
One other oddity I found and I don’t know why - the gateway backup is 125 MB.
Do you think I could up the RAM amount on the gateway/client scripts to resolve these issues? Is there a good rule of thumb? Also, is there any way to see more specifically what is causing the CPU spikes?