I've recently found a few doppelganger computers on my account:

http://www.gpugrid.net/hosts_user.php?sort=rpc_time&rev=0&show_all=1&userid=94527

Computer 143481 is the only one with credit associated with it. Any ideas on how/why these others appeared on my account? I've never detached from the project or anything along those lines. Has anyone else had this issue? It may be related to the topic I started here:

http://www.gpugrid.net/forum_thread.php?id=3497

Apologies if this is considered a double post. I wasn't sure if this is a separate topic or not.

You never reinstalled the machine or set it to an earlier state?

---

DSKAG Austria Research Team: http://www.research.dskag.at

Did you recently upgrade your hardware or software? Have you changed the names of any of your computers? Those are some of the things that can cause duplicates in your account, there's a merge button at the bottom of the page that shows your machines in your account to merge the duplicate computers (you have to be logged in).

@dskagcommunity: No, I've never done either of those.
@flashawk: In the time this computer has been online, I think I've upgraded the BOINC client once to 7.0.64 but there have been no other software changes I can think of that would affect this. I've made no hardware changes to the computer. I've never changed the computer's name either.

I attempted to merge the records, but it came back saying that two of them were incompatible. I'm not sure what that means considering they are records of the same computer. I did manage to get the number down from four to two, however.

The tasks associated with the remaining duplicate are the tasks I was having problems with described in the link I placed in my first post - the WUs that completed successfully on my computer but were "locked" somehow during upload. When they did eventually complete uploading several hours or a day later I never received credit for them because the project wasn't displaying that I had actually completed them.

I've recently found a few doppelganger computers on my account:

http://www.gpugrid.net/hosts_user.php?sort=rpc_time&rev=0&show_all=1&userid=94527

...and I recently found a doppelganger Thread - this one!

http://www.gpugrid.net/forum_thread.php?id=3497

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

I appreciate the help, everyone. I'll see if I can fix the issue with your suggestions. Again, I apologize for double-posting but I wasn't sure if this was really caused by the same problem or not.

We can see that the "ghost host" was created on 29th of september.. but seeing that every single information displayed about them is identical, I have no idea why the BOINC server software would consider them incompatible.

MrS

---

Scanning for our furry friends since Jan 2002

We can see that the "ghost host" was created on 29th of september.. but seeing that every single information displayed about them is identical, I have no idea why the BOINC server software would consider them incompatible.

MrS

If Matt was hitting a communications problem between client and server on 29 September - see my comment in the doppelganger thread - it is possible that the 'ghost host' was deliberately created by the server detecting an inconsistency in the RPC sequence number. That's a deliberate anti-cheating mechanism in the BOINC framework (designed to prevent people faking an identical HostID on multiple chunks of ironmongery, to try and boost the credit or RAC assigned to that host ID number).

I'm not for a moment suggesting that Matt was doing any such thing, but occasionally the mechanism can be tripped by some completely innocent glitch.

Since the 'ghost host' 159518 appears only to have contacted the server once, and the tasks it was allocated have now timed out, I'd advise Matt not to worry and simply carry on crunching with his 'real' host 143481. Once the ghost tasks have been purged from the database, he will be able to delete the ghost host as if it had never appeared.

It's just one of those things. Eddie's in the space-time continuum, again.

Thanks, Richard, for the thorough explanations for what may be going on here and on the other thread. Glad to know no one else is having this issue so it's not a problem with the project. I'll just carry on as normal.