Hello Everyone!! I'm not sure what to say here but, I've got a problem and I'm not sure how to fix it. A couple weeks ago, the clock on my computer got messed up. It jumped ahead about six months. I didn't catch this for about a day and a half. During that time period, work was reported on a few projects. I reset the clock/date as appropriate when I discovered the problem. After that, I got a message that additional work units would not be requested as my machine is on 100% but, that Boinc was on only 2%. I have been participating in 5 projects... working being completed on all of them each day. Now, unless I intervene, the computer only does SETI work units... and doesn't bother to download work from any of the other projects. If I suspend SETI, the computer will download from Spinhenge. If I suspend that after it downloads, the computer will download work from, the next project etc.... That's the only way I've been able to complete work on these other projects the past few days. HELP!! Do I just need to unload/delete BOINC from my computer and download a fresh copy? I look forward to hearing your suggestions. Thanks, Jeff

I tried resetting the projects... and it didn't seem to help matters or change anything.

I tried resetting the projects... and it didn't seem to help matters or change anything.

Resetting a project is a drastic action that is rarely needed and wouldn't help in this situation.

All you need to do is:-

• Stop BOINC and confirm it has stopped
  
• In your BOINC folder, find the file "client_state.xml" and open it with a text editor (notepad is fine)
  
• Find the variable called <on_frac> in the <time_stats> section
  
• Change its value (probably around 0.02...) to exactly 1.000000
  
• Save the file and exit notepad (use "save" and not "save as...")
  
• Restart BOINC
  

Your problem should be fixed. Note: The above assumes you know how to edit files with notepad. If you don't and need more help - please ask.

---

Cheers,
Gary.

There might be timestamps (in the same XML file) which need to be modified as well. If you can pull client_state.xml out of a backup from before the clock was messed up then that may help.

---

....If you can pull client_state.xml out of a backup from before the clock was messed up then that may help.

I wouldn't be game to do that as the state file contains details of all work currently on the system. If you went to an earlier backup, the list of work files would all be wrong (ie older work has been reported and removed from the state file) and you would lose all knowledge of files downloaded since the backup was taken.

If he has in fact done a full reset, he has thrown away all previous work and executables and has downloaded everything fresh. This was done after the clock was corrected so all current timestamps should be OK. He just needs to correct BOINCs erroneous view of the fraction of time that his computer is switched on. He could even be more realistic than setting it to 1.000000. If he knew he normally used his computer for 12 hours per day, for example, he could set on_frac to 0.500000 rather than the 24/7 value of 1.000000.

---

Cheers,
Gary.

....If you can pull client_state.xml out of a backup from before the clock was messed up then that may help.

I wouldn't be game to do that as the state file contains details of all work currently on the system. If you went to an earlier backup, the list of work files would all be wrong (ie older work has been reported and removed from the state file) and you would lose all knowledge of files downloaded since the backup was taken.

There's another thing: most probably all of the projects will assign new host ID to this particular machine. This is triggered due to the value of rpc_seqno on client which is supposed to be greater (or equal) to Number of times client has contacted server as recorded on project server. If the project server sees smaller number in client request, it figures that something went wrong and assigns new host ID.

An user then has to merge the two hosts ...

---

Metod ...

Hello Everyone!!

I tried every solution that was suggested. Nothing worked. Nothing cured the problem. I tried many other things... and nothing worked. The problem continued. I finally gave up. I set all of the projects to not request any new work units. Once the last project was finished... I uninstalled Boinc and then deleted it from my system. After rebooting... I did a fresh install of Boinc. And then attached my projects and everything is working perfectly. I appreciate everyone's suggestions. My computer is on 24/7 and I want it doing something always, regardless of where I am or what I may be doing. If I'm going to be paying the electric company for the juice... someone ought to benefit somewhere besides the executives at the electric company. Hopefully, my computer's efforts won't be wasted and hopefully something will come from it's numbers crunching efforts. Have a great new year everyone!!

Jeff --

Glad you got your machine straightened out, and a Happy New Year to you too ....

---

If I've lived this long, I gotta be that old

I see that Jeff managed to get his problem fixed a couple of weeks ago but the action he took was perhaps more drastic than it needed to be.

Having experienced the same problem myself at about the same time, I can state with confidence that his problem was with his long term debt.

There are several ways he could have handled this short of a complete reinstall.

He could have, as was mentioned (but not specifically recommended), used a third party utility to reset his long term debt. I'm told there are several of them out there that will do this but have never used one myself. (Note: Boinc should not be running when this is done.)

He could have edited client_state.xml manually with notepad or another editor, replacing the existing values for long term debt for all of his projects with 0.000000. (Note: Boinc should not be running when this is done.)

As he stated he did, he could have set "No new tasks" for each project and waited for the queue to drain of all work. Instead of the uninstall, reboot and reinstall of Boinc, he could simply have detached from all of his projects and then reattached to them all. This would also have cleared his long term debt and, for those who are not comfortable messing with the client_state.xml file or are unsure of how to shut down Boinc and confirm it is not running, is a completely safe alternative.

Given enough time (I would guess 6 months, maybe less, depending on how the long term debt was distributed across projects), the problem would have cleared itself up since SETI would have been 'repaid' the time debt it was 'owed' by the other projects due to the system time problem.

...I can state with confidence that his problem was with his long term debt....

It's certainly true that the stuffup with his PC's clock & date may well have screwed up LTD. However this is not the problem that was preventing the downloading of new work. There would have been at least one project (and probably more), favoured by a positive LTD that should have been able to get plenty of work.

However the error message from his original post was:-

...additional work units would not be requested as my machine is on 100% but, that Boinc was on only 2%....

Notice the 2%. His state file obviously had an active_frac whose value was 0.02 and this would have continued to haunt him even after the debts were reset. He really would have needed to reset this value to 1.000000 at the same time as the LTD values were zeroed.

I responded to him at the time but I've just now noticed why the advice wouldn't have helped him (for which I apologise). I erroneously described the parameter as on_frac rather than active_frac and the quoted message already tells us that the on_frac was already set to 1.000000 ie, 100%. on_frac and active_frac are right next to each other in the state file so, in my defense, I can say that the OP should have been smart enough to notice my mistake anyway or at least mention the discrepancy so that an immediate correction could have been made.

---

Cheers,
Gary.