| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Retired Forums
Forum: Member-to-Member Support [Read Only]
Thread: Force Server Connection for Broken Profile? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I could kick myself....
... 2 out of my 4 machines are paying the price for me not reading the definitions for execution and connection times in the profile settings and so now only run for an hour a day (i was trying to set my profile so that it gave the machines a couple of hours off processing). Is there any way I can force these machines to connect to the server and refresh thier profile settings, rather than wait for the current units to complete (at an hour a day....  )I suspect I know the answer, but it never hurts to ask... Cheers Andrew |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You could try un-installing and then try re-installing it on one machine, if this works you could then do the other
Just a suggestion Regards |
||
|
|
joatmon
Senior Cruncher Joined: Nov 17, 2004 Post Count: 185 Status: Offline Project Badges: 
|
I think they only connect to the server for profile updates between work units
---------------------------------------- EDIT - see lawrencehardin's post below for an easy way \/ [Edit 1 times, last edit by joatmon at Jan 6, 2005 1:22:46 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Use Task Manager to kill WCGridRosetta.exe. I think that will kill your current work unit and get a new work unit from the server with your new profile. Pity to lose any work, but 1 hour a day is not going to work either.
|
|||
|
||||
|
joatmon
Senior Cruncher Joined: Nov 17, 2004 Post Count: 185 Status: Offline Project Badges:
|
I just tried killing the WCGrid_Rosetta.exe using task manager, and sure enough the agent downloaded a new workunit. (I lost about one hour's crunch on the previous w/u)
so Lawrence, you were right and that should work a whole lot easier than what I had originally proposed (which I edited out to reduce confusion  )I just dumped the workunit I had been crunching, How long before WCG server realizes that the result is overdue and re-issues it to someone, or is the server smart enough to track which work units are being worked by which devices, and since I asked for a new one without returning the results from the one it knew I had before, it re-issues it sooner? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi joatmon,
Oh, I feel sure that the server software knows all this. The fun part is probably trying to optimize a response when UDMonitor might be downloading multiple work units. And of course work units lost during reinstalls probably do cause the server to wait up to 3 weeks. Remember how just a few non-converging work units started spreading to everybody back in November when a bug in Rosetta kept the program from communicating the information properly to the server? The agent would just request another Work Unit. So the server does keep close track of all this. Lawrence |
|||
|
||||
|
|