Silktest 5.5.3 locking up machine in distributed testing
I'm attempting to have a script executing on one machine start other tests running on other client machines. The tests themselves are encapsulated fully in functions. The main script has a list of string that holds the client machine names. It then for-loops through the list like so:
LIST of STRING lsMachines
case 1 // if only one machine in list
... run all test functions
case 2 // if 2 machines available...
case 1 // run a few on one machine
case 2 // and the rest on the other
The idea behind it is that I don't know how many open machines I have to test on, so a meta process finds as many free machines as possible and passes the names of the open machines into this script, which then deploys the test cases appropriately depending.
So, the problem is that when I run the script, it connects to the clients just fine, and the app's MainWin is invoked on each of the client machines, but then the host machine's CPU usage spikes to 100% and all execution stops. It's pretty severe, too. I have to kill the partner.exe process from the Task Manager in Windows.
For your info, I'm running on Win2K with Silk 5.5.3. I am using an external extension kit with Silk for this app. There are no network problems that I am aware of (i.e., each machine can see each other in every respect). The one error get before the machine locks up is a "Class not registered" error from the extension kit, which I've never seen in this context. I'm going to have my extension kit engineer look into it, but I thought I'd ask here to see if there are any bugs in the spawn/rendezvous functions that I wasn't aware of.
By the way, I googled for a Silktest issues list or something of that kind and couldn't find anything, and I could not find a list of known issues for each release on Segue's site. Could someone direct me to such a list or site, if it exists? Thanks very much.
"What we elect to call imagination is mere combination of things not heretofore combined." - Frank Norris