I keep getting the following in my message log:

Message from server: Completed result xxx refused: result already reported as success.

Where "xxx" is a work unit that was completed.

There are four tasks altogether that were completed and are shown in my tasks as 100.000% completed.

In addition it looks like on the web that I have been assigned nearly 80 tasks -- yet none of these appear in my tasks list as work to process.

This happened before and not being able to figure it out, I tried everything to clear my account -- including detaching from the project and deleting and reinstalling BOINC.

I thought that worked, but apparently not. I am still supposedly getting assigned tasks that aren't showing in my BOINC Manager and that are not getting worked on and the tasks I have already are continually trying to be re-uploaded to your server.

Any explanations? How can I clear this problem?

Thanks.

PS: I just suspended my activity until I can resolve this.

---

I keep getting the following in my message log:

Message from server: Completed result xxx refused: result already reported as success.

Where "xxx" is a work unit that was completed.

That just means your computer reported the work as complete but the acknowledgement from the server did not come back to you, so your computer thought the report had failed. Your computer then tried to report the result again but got told it was now a duplication.

The error can be caused either by problems with the server or something interrupting communication between the server and your BOINC software.

There are four tasks altogether that were completed and are shown in my tasks as 100.000% completed.

Those should be the tasks that your error relates to.

In addition it looks like on the web that I have been assigned nearly 80 tasks -- yet none of these appear in my tasks list as work to process.

This happened before...

This may be caused by something interrupting your communication with the server. Your computer asks for work but never receives it and then keeps asking the server for more work until it receives some. As far as the server is concerned you received all 80 tasks but most have got lost somewhere along the way.

The time stamps on the tasks on your account indicate that you are asking for 5 or 6 tasks at a time, then waiting some minutes/hours before trying again. That sounds like classic BOINC behaviour when your computer cannot get a response from the server - keep trying until a response is received. In your case though the server is receiving your requests and acting on them, but your computer is not getting the response.

Are you using a firewall or anti-virus/anti-spyware software? Some anti-virus programmes give false-positive readings for Rosetta files and quarantine them on arrival at your computer. Most anti-virus software maintain logs of the actions taken, so you should see quite quickly if that is where the problem lies.

I just wanted to point out that the 4 tasks were "already reported as success" by YOU. So it's not a problem that is causing any harm here. When you detach from a project, you abort all of the existing work units, they are going to show on the website until they age and are removed by the server. Those relate to work you never ran, so again, no harm there.

Once your machine sees that the tasks have already been reported, the BOINC Manager should clear them from the list, that is all the confirmation it should need. Is that not happening? (If so, that's something that would need to be reported through Berkeley)

As Murasaki explains, if this sort of thing is happening often, it is generally an indication of communications problems between your machine and the server. The server responds to your request for work and hands you 6 tasks, but the transmission is cut-off before your machine receives the end tags (if this occurs a message is issued in the log on your machine) and so the server reply is lost. Those 6 tasks eventually reach their 10 day deadlines and are issued to someone else. Then they age and eventually are removed from the list of tasks.

So, no need for concern over the existing tasks. The concern should be whether this is an ongoing problem when BOINC is running normally (i.e. no detach or aborts done). If so, a careful review of your network, firewall and anti-virus activity may turn up ways to avoid it.

---

Rosetta Moderator: Mod.Sense

It looks like dperreault has been completing and returning every task received since 16th December, so whatever the problem was appears to have been resolved.

Hi. I apologize for not responding sooner. I didn't realize someone had replied as I hadn't "subscribed" to the thread. I assumed this would have happened automatically as the originator.

Anyway, I've done that now.

I realize in the long run the only harm being done is that my machine is requesting tasks that aren't getting assigned to someone else while your server is waiting for me. Sure, they'll get assigned to someone else, but it'll be later rather than sooner. So it is something I need to resolve.

The thing is, I've been running SETI for years. I only starting running the Rosetta project because SETI was having some issues with a bad server and I figured I might as well attach to another project while I was waiting for them to get things running again. I've never had this problem with SETI, so I'd find it odd to be an issue with my firewall or antivirus, though I suppose it's a possibility. At present I'm running Win 7 with the Win 7 firewall and Microsoft Security Essentials (MSSE) for anti-virus. BOINC is set as allowed through the firewall. I've never had a false positive with MSSE and BOINC, but just to be sure, I just marked the BOINC data directory to be excluded as a scan location in MSSE. I already had the BOINC program directory excluded. BOINC and BOINC Manager are both marked as allowed through the firewall and as I mentioned, everything works fine with the SETI project so I don't think this is the issue anyway.

Things were going OK, but the problem returned on the 22nd. I now have 12 tasks that have previously been uploaded that haven't cleared.

I do see one of them reports that it was reported as an error. I wonder if that had anything to do with this issue?

So, is there an easy way to get these 12 tasks out of my queue so that I can start processing new ones? It looks like I actually have more tasks in the program data directory.

Any completed tasks that remain on your machine are not holding up getting new work.

BOINC server code allows the project to configure whether it should double check everything you report and see if there are tasks that the server thinks you have, that you don't have on your machine. This takes considerably more processing time for every scheduler request to the server. R@h does not recover lost tasks.

Where are you saying you see 12 tasks that don't belong? On the R@h website for your machine? Or on the BOINC Manager display on your machine? From the website, it appears your machine has dozens of tasks (ok 100s actually). Indeed about 100 tasks per day have been requested and sent to your machine, you shouldn't have a lack of work here. Are you now seeing 100s of tasks ready to run in your BOINC Manager display?

---

Rosetta Moderator: Mod.Sense

Any completed tasks that remain on your machine are not holding up getting new work.

Yes, I understand that. That's not the concern. The concern is that they remain in my task list and are filling up my log file.

Where are you saying you see 12 tasks that don't belong? On the R@h website for your machine? Or on the BOINC Manager display on your machine? From the website, it appears your machine has dozens of tasks (ok 100s actually). Indeed about 100 tasks per day have been requested and sent to your machine, you shouldn't have a lack of work here. Are you now seeing 100s of tasks ready to run in your BOINC Manager display?

I am seeing them on my machine -- in the BONC task manager. They just sit there listed as "ready to report."

Re the 100's of tasks -- unfortunately I wish that were so. My tasks in BOINC show ONLY the 12 that are "ready to report." Every time my machine tries to send these to you, the response is that I have already sent them, BUT I am then supposedly sent several more in return. It's just that I never get them.

THAT is my concern. I supposedly have 100's of tasks to be processed (that no one else can process until my time to report expires) yet on my machine there's nothing to do.

I have NEVER had this problem with SETI in the years I've used BOINC (back to the old version right up until now) so I am doubtful it is my firewall or antivirus.

--Doug

OK, so the reissuing of lost tasks is not the concern here.

After the scheduler request (update to project) with R@h completes, what (other) messages appear in the messages tab of the advanced view? Do you see anything about no end tag found? Or how many tasks were received?

The reason we were asking about antivirus and firewalls was because these might explain why the server very consistently thinks it has sent something to your machine, and yet your machine doesn't seem to have received it... or received it and then destroyed it (as if it were a virus).

---

Rosetta Moderator: Mod.Sense

As fate would have it, I can't tell you exactly as I "aborted" the completed tasks and after a while those tasks finally disappeared from my tasks list. Now the message log has no errors and everything is working correctly again.

I don't recall seeing anything about "end task" messages.

What I would see is a message that it was reporting tasks as complete and requesting new tasks. Then the server would respond that the tasks were already reported, but then it would send me new tasks (around 6 I think). Then a few minutes later, the process repeated. The new tasks I was supposedly sent, however, never showed up in my tasks list.

As I mentioned, I've been running BOINC with the SETI project for years and never had this type of problem. They were down for a while, which is why I searched for another worthy cause. They're back up now, and I'm now successfully processing for both the SETI and Rosetta projects at the moment.

Anyway, if it happens again, I will copy the message log here. For now it's working correctly again.