Send Unix Command fails to move on when launching permanent remote dialogs

Question

Level 1

0 points

Send Unix Command fails to move on when launching permanent remote dialogs

I use ARD's Send Unix Command to "leave" a lasting dialog on machines when a task is completed, until someone manually closes the dialog. The problem is that ARD is forcing itself to wait on machines after we know we're done with them. If I had 50 target computers to run the script, ARD would probably queue most of them until the first 15 (or whatever the parallel deployment max is) have manually had the dialog closed.

My "Unix Command" for ARD is simple bash calling a java program that leaves the window on the target PC. I broke it down so you can replicate with this one-liner:

osascript -e 'tell application "System Events" to display "Message"'

Even with ampersands at the end of such a command, ARD insists to wait however long each computer has the dialog up. How could I tell ARD to let my process run indefinitely and that it's OK to just hit-and-run through my list of macs?

Thanks

Apple Remote Desktop-OTHER, Mac OS X (10.6.8)

Posted on Jun 29, 2011 5:56 AM

Reply

Answer 1

TeenTitan

Level 4

2,640 points

Jun 29, 2011 9:05 PM in response to Vlueboy

Well ARD has the send messages command. Is there a reason you want to use applescript via terminal instead of the built in messaging command?

Reply

Answer 2

Vlueboy Author

Level 1

0 points

Jun 30, 2011 3:30 PM in response to TeenTitan

Sorry about not being very clear originally. We don't actually send remote notices to the target machines via ARD. The command

osascript -e 'tell application "System Events" to display dialog "Message"'

is just an example of what ARD is having issues with for me: If the script for the target machine pushes any user-facing GUI at all, it will not quit until someone manually OKs or closes the window... no amount of messing with the command by giving it ampersands or starting subshells prevents ARD from minding every single one of the subprocesses that it has started.

There are posts here where we see that a certain task never even "ends" in ARD's eyes: the scripted reboot

A colleague has suggested that we update our application to echo command information instead of triggering an "Everything was OK" dialog. Most of the time this will fail because you do not have access to the source code in the real world. In my case, we do, but making the changes requires reviewing, changing the code and new testing for the triggered application.

Still, there is a flaw in not having an option for ARD to just hit-and-run with a script to, say, open a presentation in 100 computers and leave it running. Currently, ARD will hit a few and hang on "Running" as soon as its whatever maximum parallel workstations is stuck running this script.

ARD designers seem to have planned it "we must receive the return value of this open-ended script," while clients see it as "please let us treat the job as finished as soon as the last line has been triggered"

Reply