Still you! :-p

BOINC has an internal and external CPID. The internal ID is in the database and your client_state. The external one is on the profile page and in the exported stats.

Read the link I gave, it's all explained (badly, but it seems to be accurate).

So, you'd better edit your post before you get spoofed.... ;-)

Hi Didactylos, thanks. OK, I've connected MusicBox to Seti. It will finish a result in a day or two, and then it sounds like everything should be OK from then on.

I had been running the menu bar version of BOINC on MusicBox. I upgraded to the full GUI version as part of connecting to WCG. So, I probably managed to delete all the old account info as part of that process. On George, I didn't change BOINC clients, so I didn't erase any old data. Which is probably why George didn't become an orphan and MusicBox did.

Part of my confusion came from misreading something in http://boinc.berkeley.edu/trac/wiki/CrossProjectId. On the 3rd bullet point, "When the BOINC client makes an RPC request to a scheduling server, it scans the accounts with the same email address, finds the one with the oldest creation time, and sends the CPID stored with that account." I read this as scanning the accounts on the server, not scanning on the local machine. But, now that I know the right answer, I see that it can't scan on the server without sending the email address, which it says it doesn't do.

Still, modifying the sentence might help make this more clear: When the BOINC client makes an RPC request to a scheduling server, it scans the accounts _on that client_ with the same email address, finds the one with the oldest creation time, and sends the CPID stored with that account.

I would not have jumped to the conclusion I did, which led me down a bad path.

I appreciated the meta-discussion. I learned enough to really understand what probably happened, instead of just enough to fix the problem without really understanding. So, thanks to you and to Sekerob for this.

Pat

I think a diagram would probably help. If you like, I'd be happy to critique and suggest improvements if you want to draft something.

Pat

The HPF2 work units were shortened by quite a bit in the last week and the deadline extended from 9 to 11 days. Also WCG sets the default CPU use to 60%... might try the "Maximum Output" profile for BOINC which uses 100%. If u run the UD client which is expected to remain used until end of FA@H and HPF2, you have a 2 month deadline (3 weeks wallclock run time or 2 weeks CPU time whatever comes first). Here's the link: http://www.worldcommunitygrid.org/bg/WorldCommunityGridAgent.exe

and the Throttle tool to go with UD (realtime control from 25 to 100%): http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=2683