See Message 993241 by Aurora Borealis:

Boinc uploads the results immediately on completion if the Internet is available. It tries to bundle the actual reporting of completed work to ease the load on the server. Accessing the database to report a group of results only adds a small percentage of overhead more than an individual report. Boinc also ensures that work is not reported late by following a set of reporting rules as previously posted by JM7

BOINC returns work to a project at the first of:

1) 24 hours before deadline.
2) Connect Every X before deadline.
3) 24 hours after completion.
4) 60 seconds after the completion of upload if upload completes after any of 1 to 3.
5) On a work request to that project.
6) With any other reported work for that project.
7) On a trickle up to the project (CPDN only as far as I know).
8) On a trickle down request (no projects use this as far as I know).
9) On a server specified schedule (some project uses this, but not certain which).
10) On a BOINC Account Manager specified connection (only if using a BOINC Account Manager).
11) When the user pushes the update button.

Rules 3, 5 and 6 are the ones that are most commonly seen being used especially if you are attached to more than one project.

Gruß,
Gundolf

Hellsheep try this in a cc_config file


<cc_config>
<options>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>

should help

---

I have 4 pending units on my Linux box, no GPU. which is an average CPU, and 5 pending units on my virtual OpenSolaris machine, which is 4 times slower!
Tullio

---

Hellsheep try this in a cc_config file


<cc_config>
<options>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>

should help

Thanks, i'll try that.

See Message 993241 by Aurora Borealis:

Boinc uploads the results immediately on completion if the Internet is available. It tries to bundle the actual reporting of completed work to ease the load on the server. Accessing the database to report a group of results only adds a small percentage of overhead more than an individual report. Boinc also ensures that work is not reported late by following a set of reporting rules as previously posted by JM7

BOINC returns work to a project at the first of:

1) 24 hours before deadline.
2) Connect Every X before deadline.
3) 24 hours after completion.
4) 60 seconds after the completion of upload if upload completes after any of 1 to 3.
5) On a work request to that project.
6) With any other reported work for that project.
7) On a trickle up to the project (CPDN only as far as I know).
8) On a trickle down request (no projects use this as far as I know).
9) On a server specified schedule (some project uses this, but not certain which).
10) On a BOINC Account Manager specified connection (only if using a BOINC Account Manager).
11) When the user pushes the update button.


Rules 3, 5 and 6 are the ones that are most commonly seen being used especially if you are attached to more than one project.


Gruß,
Gundolf

I am aware of that post, the problem is my results went through the whole weekend last week without reporting anything by itself, i had to report them manually.

---

- Jarryd

I am aware of that post, the problem is my results went through the whole weekend last week without reporting anything by itself, i had to report them manually.

If there really are no connection errors and/or backoffs, that sounds like a BOINC client error (not SETI).

I'm not quite sure which log flags to enable to track this down.

Gruß,
Gundolf

I am aware of that post, the problem is my results went through the whole weekend last week without reporting anything by itself, i had to report them manually.

BOINC should never wait more than 24 hours to report completed tasks. Are you sure that Network Activity is enabled? If you click 'update' to report them manually, BOINC temporarily opens the network to comply with your request, but them shuts it off again.

as per other post ... pending is now at 493k and rising ... normally about 130-140k .. this has occurred since the changes ...

[noticed effect just after we went to new methodolgy]

---

I am aware of that post, the problem is my results went through the whole weekend last week without reporting anything by itself, i had to report them manually.

BOINC should never wait more than 24 hours to report completed tasks. Are you sure that Network Activity is enabled? If you click 'update' to report them manually, BOINC temporarily opens the network to comply with your request, but them shuts it off again.

Network activity was on.

Today after my rig finished crunching all its work i detached to clear our 400+ ghost workunits.

Now that they're gone and i've set it all back up everything works perfect. I've gotten another 1300 work units and its reporting after every task or two.

---

- Jarryd

310,000 pending :D

That's 17.85 times my RAC :P

---

Rig name: "x6Crunchy"
OS: Win 7 x64
MB: Asus M4N98TD EVO
CPU: AMD X6 1055T 2.8(1,2v)
GPU: 2x Asus GTX560ti
Member of: Dutch Power Cows

This morning before we went online I had 378,492.63 pending and now I have 499,615.29

---

Official Abuser of Boinc Buttons...
And no good credit hound!

This morning before we went online I had 378,492.63 pending and now I have 499,615.29

It just dropped from 310,000 tot 250,000 :P

---

Rig name: "x6Crunchy"
OS: Win 7 x64
MB: Asus M4N98TD EVO
CPU: AMD X6 1055T 2.8(1,2v)
GPU: 2x Asus GTX560ti
Member of: Dutch Power Cows

Bearcat,
The only thing it messes up is the statistics tab on your BOINC Manager. There is a way around this but I will let someone else explain that, I just live with it myself. If you set to no new tasks, run down the work you have left, report your finished work, then hit detach, when you reattach to SETI it will mark all the ghosts (or any work you didn't finish) as client detached and will issue you new work. You do not lose any total work or RAC, you just pick up with the new work and the ghosts are reissued to another cruncher.

Hi, I tried a detach to get rid of the GHOST WorkUnits.
That did not release them..
Example:

http://setiathome.berkeley.edu/workunit.php?wuid=629999178
and the computer ID of 52348ee. It still shows assigned after 20 hours.

It makes sense that detach does a return of WUs that are on your machine -
but not the ghosts.
The ghost units would still appear in the SETI database - until they timeout.
I'll look for more details in the ghost topic...

Jay

Bearcat,
The only thing it messes up is the statistics tab on your BOINC Manager. There is a way around this but I will let someone else explain that, I just live with it myself. If you set to no new tasks, run down the work you have left, report your finished work, then hit detach, when you reattach to SETI it will mark all the ghosts (or any work you didn't finish) as client detached and will issue you new work. You do not lose any total work or RAC, you just pick up with the new work and the ghosts are reissued to another cruncher.

Hi, I tried a detach to get rid of the GHOST WorkUnits.
That did not release them..
Example:

http://setiathome.berkeley.edu/workunit.php?wuid=629999178
and the computer ID of 52348ee. It still shows assigned after 20 hours.

It makes sense that detach does a return of WUs that are on your machine -
but not the ghosts.
The ghost units would still appear in the SETI database - until they timeout.
I'll look for more details in the ghost topic...

Jay

What you actually need to do is to re-attach to the project, following the detach. The actual process of detaching doesn't communicate anything to the project, just cleans the files from your machine. Re-attaching does communicate back to the project, requesting a host ID - you should get the same one as before - and that's the moment when the database gets cleaned up.

Bearcat,
The only thing it messes up is the statistics tab on your BOINC Manager. There is a way around this but I will let someone else explain that, I just live with it myself. If you set to no new tasks, run down the work you have left, report your finished work, then hit detach, when you reattach to SETI it will mark all the ghosts (or any work you didn't finish) as client detached and will issue you new work. You do not lose any total work or RAC, you just pick up with the new work and the ghosts are reissued to another cruncher.

Hi, I tried a detach to get rid of the GHOST WorkUnits.
That did not release them..
Example:

http://setiathome.berkeley.edu/workunit.php?wuid=629999178
and the computer ID of 52348ee. It still shows assigned after 20 hours.

It makes sense that detach does a return of WUs that are on your machine -
but not the ghosts.
The ghost units would still appear in the SETI database - until they timeout.
I'll look for more details in the ghost topic...

Jay

What you actually need to do is to re-attach to the project, following the detach. The actual process of detaching doesn't communicate anything to the project, just cleans the files from your machine. Re-attaching does communicate back to the project, requesting a host ID - you should get the same one as before - and that's the moment when the database gets cleaned up.

I think I'll just stick with the ghost units for the moment. I don't seem to have any shortage of work on my comps and the ghosts I have will eventually return to the pool by themselves. Good to know that there is an emergency procedure just in case. Thanks guys.

---

New record for me....
Pending Cedit: 529,024.49

---

Official Abuser of Boinc Buttons...
And no good credit hound!

I am at 246,393.75 ... that's over 12 times my RAC. I haven't checked for a while, but wasn't it 4-5 times ? Hard to believe this is caused by ghosts.

---

The "ghost work units" exist only on the client computer work unit listings at Berkeley. They therefore cannot influence either "pending" or "RAC" by the very fact that they have not and will not be crunched.

The higher ratios are probably caused by fustration with the 3 day outage and people dropping out of the project, abandoning their work.

---

Boinc....Boinc....Boinc....Boinc....

The "ghost work units" exist only on the client computer work unit listings at Berkeley. They therefore cannot influence either "pending" or "RAC" by the very fact that they have not and will not be crunched.

The higher ratios are probably caused by fustration with the 3 day outage and people dropping out of the project, abandoning their work.

I don't think you are entirely correct. Ghost work units will increase the number of pending results, when people don't use the detach-reattach method to release them so they can be assigned to other clients. This means a number of results will remain pending for a longer than normal time, because the server thinks they are being processed while no client actually has the unit to work on. However, I do not believe that ghost units would have such an impact on the pending results. Or am I misunderstanding ghost units entirely ?

---

Not quite right Geek, if the ghosts are showing as on your wingman's computer you will be stuck in pending until either they are aborted by a detach/attach or timed out. I do agree a lot of crunchers are dropping out though. It's sad really but not much you can do.

Bearcat, ghosts do have a big impact but a lot of it is people slow to get their work in before the next outage. You will see a big jump in pendings once you report because your wingmen haven't gotten theirs in yet. It will fall as the days go by but again, a lot just don't make it in.

---

PROUD MEMBER OF Team Starfire World BOINC

I don't know where your getting your list of "pending" but I am getting mine from the "Pending credit" under my account listing.

[edit] and work must be crunched to appear on that listing.

---

Boinc....Boinc....Boinc....Boinc....