Posted by Barbara Brazil on August 26, 2016 at 13:59:20:
In Reply to: COMET MAXES OUT CPU AVAILABLE MEM posted by HAL MARCH on August 26, 2016 at 13:31:56:
Hal:
We have only seen this in Comet32 clients with one single-core CPU, which I doubt is the case with your server. It can happen if a program has a tight loop of code. That same program run on a multi-cpu system will distribute the load amongst all processors, none of which will approach 100%.
We found this issue with one of our own utility programs, the "forever mode" print feature of DocMgr. We solved the problem by adding a PAUSE 2 at the top of the loop of code that was checking to see if any documents were waiting to be printed. This change was released in REL 16.01.
Finding a similar loop of code in your own application can be tricky since once the CPU hits 100% it's hard to do anything. Perhaps polling the users to see what they were doing at that time will help identify the program.
Comet32 relies on Windows to do CPU management. Comet16 had its own time-slicing algorithm to share the processor. We are studying where to perhaps add task break logic to Comet32 to prevent this, but again we've only seen it on clients with one single-core CPU.
Each file can be a maximum of 1MB in length Uploaded files will be purged from the server on a regular basis.