Posts by Bernhard Frey

1) Message boards : SZTAKI Desktop Grid : Something a bit odd here ... (Message 5144)
Posted 3961 days ago by Bernhard Frey
I had this problem too. And I have still some WU\'s with 5 or 6 returned results and validator in state checked, but no consensus yet.

My guess is the following:
There have been lots of problems with the earlier 2.0x versions in the different OS-Systems (MAC <--> Win <--> linux). So I suppose that it is often just a game of chance, who is the lucky one to be selected as valid one. We\'ve seen this before restart of SZTAKI, that different OS-Systemts or app-versions had a different character (LF/CR) at the and of the result file.
Furthermore there are open threads reporting an empty output problem and an issue that input lines are skipped, when boinc client is restarted.

So there is much space for speculating which concret issue did hit you. Fact is, that there are still some open issues that cause invalid results. This project seems to have a very low staffing, but I hope that Adam will mange to solve this issue, as he did before for other problems. It took time, but he manged it.
Until this we just can hope, that he finds time to grant credits for those problem results, before the automatic deleter has purged them and so covered all the tracks.

Bernhard
2) Message boards : SZTAKI Desktop Grid : WU 2515 sucess but 0.0 credit granted? Why? (Message 4753)
Posted 3986 days ago by Bernhard Frey
Anotherone with Error to many results:
wuid=6128

@Adam
Are there any plans to solve that problem (e.g. enlarge the settings, ganting credit manually].
3) Message boards : SZTAKI Desktop Grid : Workunit error - check skipped (Message 4585)
Posted 4005 days ago by Bernhard Frey
Let\'s hope Adam will manually grant credit for those WUs with Too many total results. And let\'s hope he does it quickly, because the automatic delete fueature will purge those WUs (like all other finished WUs) out of the database in a few days.
4) Message boards : SZTAKI Desktop Grid : Ad@m\'s POD (page 7) (Message 3679)
Posted 4105 days ago by Bernhard Frey
Good luck, XJR-Maniac,

I had had also a >300 credit WU that lost when it was validated against 2 Linix boxes. In the meanwhile the WU was deleted from the database (one week after the validator found one of the linix result as the \"correct\" canonical one).

And my credit: blown with the wind.

So try to fight, befor your WU is deleted (what will happen in one week, cause ther is a valid canonical result)

5) Message boards : SZTAKI Desktop Grid : New applications 2.0 /2.03 online - Feedback (Message 3345)
Posted 4127 days ago by Bernhard Frey
got one new WU with a deadline of 100h.
But when I lock on the percentage done, it seems it would take over 70h to complete.
Suspended it for the moment. May be some others had the same experiences.
6) Message boards : SZTAKI Desktop Grid : New applications 1.02 / 1.03 online - Feedback (Message 2366)
Posted 4225 days ago by Bernhard Frey
The restart/queckpoint problem is still unsolved in the new windows version 1.02 of the application.

If one stops the boinic core client (e.g. to reboot the computer), the CPU time of a partialy completed WU starts at zero (the percentage complete continues at the point it was). All those restarted results are invalid.
I had 3 of them last 2 days:

WU 425721
WU 427689
WU 428641

As long as restarting doesn\'t work propably, we should return to shorter WU\'s (calc time 30 to 60 minutes).
7) Message boards : Észrevételek, tapasztalatok : A problem : wus sent 8/9 march (Message 2310)
Posted 4230 days ago by Bernhard Frey
some more details about the same problem as posted here
8) Message boards : Észrevételek, tapasztalatok : Results without their WUs!? (Message 2309)
Posted 4230 days ago by Bernhard Frey
Some more aspects on this problem. I guess it started during th DB-maintenance on 8th of march.

1) Results without their WUs
I\'ve got about 10 WU\'s from 27th and 28th of February on my host list, where
a restult is still in the database, but the corresponding workunit isn\'t found.
All are workunits aborted via GUI RPC by myself (that was during the timeframe, when the client downloaded lot\'s of work with an estimate of 15 minutes, but the took ~3h)
Examples:
Result 1073171
Result 1073211
I\'m afraid those results won\'t be automaticly deleted.

2) WU\'s where some of the results have been deleted
There are WU\'s that had had a quorum (out of 3 relults) and credit was granded. All results, but the canonical result has been deleted. Then the 4th WU came in time, but later and now this late one is pending. The validader takes no notice of them.
The stay there with two remaining results (the canonical one validated and with granded and the late in pending state)
Examples:
a) I\'m the one who was late and have pending resultes (about 50 in my host list)
WU 342425
WU 342426
b) I hold the canonical result and somone else was late
WU 350938
WU 350938


Home | My Account | Message Boards


Copyright © 2017 SZTAKI Desktop Grid