two million hours to completion

gravywavy
gravywavy
Joined: 22 Jan 05
Posts: 392
Credit: 68962
RAC: 0
Topic 189128

more of an oddity than a problem, but this machine just had a power outage for about three hours, and when I rebooted BOINC came back and restarted the current wu, showing 9 hours already processed and over two million hours to go...

several days got chipped off the estimate every second and it is gradually getting more reasonable. After just over an hour it is down to an estimate of 300 hours and decreasing by about 10 min each second.

The progress bar is also showing a much smaller % progress that I'd expect, but consistent with the estimated time to complete.

Presumably the software is not treating the gap in processing properly, but that is strange as I am sure I've had outages before and not had this side effect...

I've reported it partly for the amusement value and partly in case this behaviour leads the technical team to a more serious bug - this may be irrelevant however as I am still running boinc v 4.19 and you may not care about bugs in old versions.

~~gravywavy

history
history
Joined: 22 Jan 05
Posts: 127
Credit: 7573923
RAC: 0

two million hours to completion

Gravy; I've had the same behavior with 4.19. The app will start processing the WU over, and will usually take longer than normal for that WU, given the average for that rig.

gravywavy
gravywavy
Joined: 22 Jan 05
Posts: 392
Credit: 68962
RAC: 0

[...] The app will start

Message 11424 in response to message 11423


[...]
The app will start processing the WU over
[...]

good thought tweakster,

the timings are consistent with the wu starting from the beginning again but the gui not resetting the cpu used clock. Estimate of two million hours makes more sense if it was figuring it had made

~~gravywavy

history
history
Joined: 22 Jan 05
Posts: 127
Credit: 7573923
RAC: 0

gravy; We bust our humps

Message 11425 in response to message 11424

gravy; We bust our humps putting together a killer rig or two and we get some software spliced together by Miss peach's fourth grade class. I am having no joy with the "quota 8" bug right now. If I could just delete the rigs in my computer listing that are artifacts, I'd be a happy camper. But NOOOOO, we have to merge them, eh? What a software labryth that is. Anyway, I signed on with seti on one of my rigs that Einstein loves to hate yesterday using Boinc 4.37. So far, this version seems to be reasonably well done. 4.25 is a serious piece of crap. Thanks for the post and add my appreciation for the charity of all who waste their attention span on my rant.

regards-tweakster

LichtJF
LichtJF
Joined: 14 Mar 05
Posts: 7
Credit: 13497
RAC: 0

Yes, I was having problems

Yes, I was having problems with wu's as well. Version 4.37 seems to have resolved all my problems. So Far.......


Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5877
Credit: 118573834239
RAC: 17522211

I am having no joy with the

Message 11427 in response to message 11425


I am having no joy with the "quota 8" bug right now.
....
But NOOOOO, we have to merge them, eh? What a software labryth that is.

I'm sorry but I really don't understand your "quota 8" bug as you call it. If your best cpu can do about 4-5 units per day, why is there a problem with a limit of 8???

What's the problem with merging?? It's really a breeze. I've had a few users with administrator priviledges on boxes I don't own (but have permission to run on as a service), decide to delete the whole BOINC directory because they didn't know what it was. On reinstalling, the box gets a new CPU ID. However all I do is merge the old deleted instance with the new instance and there is no problem. If it's not the same physical hardware, don't do the merge. Just be patient until all work expires from the online database and then you'll be able to delete the defunct CPU ID. Merging however retains all the previous credit that was accumulated on that box.

Cheers,
Gary.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5877
Credit: 118573834239
RAC: 17522211

.... showing 9 hours already


....
showing 9 hours already processed and over two million hours to go...

When the power went down, something was left in an inconsistent state so the processing was restarted. The bug must be that the elapsed time was left at 9 hours. The routine that estimates the finish time will obviously figure an enormous value if it thinks that 0.00% was achieved in 9 hours. A minute later when it has done 0.xx% in 9hrs 01mins, have a guess what the new finish time will look like. When the work actually finishes, this phantom 9 hrs will be added to the normal finishing time whatever it happens to be and your box will make an inflated claim for credit which just may cause you to be accused of cheating :).

Cheers,
Gary.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.