Boinc "report deadline" differs from real deadline
24 Mar 2005 19:13:12 UTC
Topic 188717
(moderation:
The "report deadline" inside Boinc Manager are all 1 hour later compared to the real deadline as seen on the website.
It looks like this an application bug and not a Boinc bug as for S@H as well as LHC@Home both deadlines are identical.
For info: my local time = UTC +1h.
One example: this WU has a reported deadline of 31 Mar 2005 17:48:09 UTC. Boinc Manager reports the deadline as 31/03/2005 19:48:09 local time, which is 18:48:09 UTC.
Is there any relationship between the difference of local time against UTC time and the difference seen in the deadline reporting?
Greetings from Belgium
Thierry
Boinc "report deadline" differs from real deadline
)
As I checked to be sure and the 60 minutes are over I need to make a correction:
the bug is NOT inside the application. For the S@H WU's I have the same difference. Which means the bug is inside Boinc apparently.
Should have check it before ;(
Greetings from Belgium
Thierry
My einstein times show
)
My einstein times show correct on my linux and windows systems with 4.25, but my seti times are off 1 hour.
I'm UTC-5 (US Eastern Time). On my linux 4.25, the times also indicate the time zone, though it doesn't on my windows systems. I see in linux that the seti units are indicating EDT, while the einstein units are indicating EST. (EST is Eastern Standard Time and EDT is Eastern Daylight Time. Daylight time in the US = Summer Time in Europe.)
If the standard versus daylight time is allowed for, my seti times would be correct, but again, my einstein times show correct as it is.
> My einstein times show
)
> My einstein times show correct on my linux and windows systems with 4.25, but
> my seti times are off 1 hour.
>
> I'm UTC-5 (US Eastern Time). On my linux 4.25, the times also indicate the
> time zone, though it doesn't on my windows systems. I see in linux that the
> seti units are indicating EDT, while the einstein units are indicating EST.
> (EST is Eastern Standard Time and EDT is Eastern Daylight Time. Daylight > time in the US = Summer Time in Europe.)
>
> If the standard versus daylight time is allowed for, my seti times would be
> correct, but again, my einstein times show correct as it is.
I am running Windows XP SP2, BOINC 4.25, in UTC-6 (US Central Time), and have the same results as Darren - BOINC E@H Report Deadline agrees with website, but S@H BOINC Report Time is 1 hour later than website.
It seems that Darren has it nailed - the US Daylight Time vs. Standard Time explanation for the S@H discrepancy looks very credible.
------------------
If I love truth, then why don't I love correction? Perhaps Jack Nicholson was right - maybe I really can't handle the truth!
> It seems that Darren has it
)
> It seems that Darren has it nailed - the US Daylight Time vs. Standard Time
> explanation for the S@H discrepancy looks very credible.
And remember that any einstein units you get through this coming Saturday in the US will be due back before Daylight Savings Time begins, but any seti units you have gotten since last Sunday will not be due back until after it begins, as it begins on April 3 in the US.
Thierry is from Belgium, where they begin Summer Time this coming Sunday, so any work units from einstein sent to anyone in Belgium since last Saturday will have a due date after Summer Time takes effect.
I would guess (but this is just a guess) that the boinc manager makes the determination of how it displays the time at the local system level, so anything that's due after a time change would account for that time change in the display based solely upon whether your local system is set to make adjustments for time changes when they occur.
> > It seems that Darren has
)
> > It seems that Darren has it nailed - the US Daylight Time vs. Standard
> Time
> > explanation for the S@H discrepancy looks very credible.
>
> And remember that any einstein units you get through this coming Saturday in
> the US will be due back before Daylight Savings Time begins, but any seti
> units you have gotten since last Sunday will not be due back until after it
> begins, as it begins on April 3 in the US.
>
> Thierry is from Belgium, where they begin Summer Time this coming Sunday, so
> any work units from einstein sent to anyone in Belgium since last Saturday
> will have a due date after Summer Time takes effect.
>
> I would guess (but this is just a guess) that the boinc manager makes the
> determination of how it displays the time at the local system level, so
> anything that's due after a time change would account for that time change in
> the display based solely upon whether your local system is set to make
> adjustments for time changes when they occur.
Hello Darren,
I was thinking about this DST (dailight saving time) afterwards.
Boinc Manager indeed display the time at local system level.
Will keep an eye on the difference between S@H and E@H after we jumped to summer time here.
Greetings from Belgium
Thierry
After jumping here in Belgium
)
After jumping here in Belgium to Summer time the difference between deadline in Boinc Manager and the deadline of reporting mentionned on the website remains: there is still one hour difference between both.
This difference of 1 hour is for E@H as well as for S@H.
Greetings from Belgium
Thierry
> After jumping here in
)
> After jumping here in Belgium to Summer time the difference between deadline
> in Boinc Manager and the deadline of reporting mentionned on the website
> remains: there is still one hour difference between both.
>
> This difference of 1 hour is for E@H as well as for S@H.
Did you reboot your machine?
This does not help the servers, as it should not affect the client. But sorry, I do not trust Microsoft when changing between normal time and Daylight Saving. People who use (or better: used) CVS/WinCVS know what I mean.
Update - update - update
I checked one Workunit: http://einsteinathome.org/task/2452575
Web Page: 3 Apr 2005 1:42:30 UTC
Boinc bash script¹): Son 03 Apr 2005 03:42:30 CEST
¹): http://www.boinc.dk/download.php?file=boinc-1.5.tar.bz2
So I do not see any difference?
Another workunit which was downloaded after the timeswitch:
http://einsteinathome.org/task/2480925
Web Page: 3 Apr 2005 8:14:54 UTC
Boinc bash script: So 03 Apr 2005 10:14:54 CEST
> After jumping here in
)
> After jumping here in Belgium to Summer time the difference between deadline
> in Boinc Manager and the deadline of reporting mentionned on the website
> remains: there is still one hour difference between both.
>
> This difference of 1 hour is for E@H as well as for S@H.
I have no idea where boinc gathers your system time info from, but the very first entry in your client_state.xml file should be your timezone, in seconds offset from UTC.
If you're going to try to track down where the difference is coming from, you may want to check that first and make sure it's got you in the right timezone.
> I have no idea where boinc
)
> I have no idea where boinc gathers your system time info from, but the very
> first entry in your client_state.xml file should be your timezone, in seconds
> offset from UTC.
>
> If you're going to try to track down where the difference is coming from, you
> may want to check that first and make sure it's got you in the right
> timezone.
Timezone in xml file is given as 3600.
After rebooting, xml file is unchanged, time is still given as 3600.
Actually, local time here is UTC + 2h. So, something is definitely not correct.
Greetings from Belgium
Thierry
As good as I can figure from
)
As good as I can figure from digging around, I think boinc calls the timezone from the OS, so it should show the same offset that your OS shows. But I have no clue if windows actually changes the offset from UTC when the time changes or if it compensates some other way.
I know very little about windows and how it thinks. If you pull up your time properties in windows, is the UTC offset correct there? If windows shows UTC+2, boinc should be showing +7200. Lastly, I would suggest checking the BIOS time against what windows shows. If windows is actually changing the time (rather than just the way it displays it), this should be reflected by comparing it to the BIOS.
Maybe someone who knows more about windows in particular can clear some of that up though.