1) Message boards : Webpage and Application Code Discussion : Invalid results in DNA@home caused by quitting and restarting BOINC client? (Message 4946)
Posted 11 Jan 2015 by Ananas
There is an issue where the checkpoint/restart generates a new seed, which causes the result to be different from the uninterrupted result. It does not happen all the time, some checkpoints do work properly.

It has been reported but afaik. the cause has not been found yet. The developers are working on it.
2) Message boards : Webpage and Application Code Discussion : Link my old DNA account created with now-orphaned email address? (Message 4929)
Posted 28 Dec 2014 by Ananas
If I understand that combine dialog right, you can just fill it out with your old account names together with your old emails and it should combine them into the account, you're currently logged in with.

p.s.: don't forget to join your team (if you had one) with your new account, because this information is most likely not part of the "combine" function.
3) Message boards : Number Crunching : Error with new longer tasks. (Message 4927)
Posted 27 Dec 2014 by Ananas
...

However I did locate the source of the validation error.
In the stderr output file there is a value "argument seed (number)" at the start and a value "seeding (number)" at the end.

If these are equal then the result is valid, unequal is invalid.

It looks to me that restarting from a checkpoint can cause this seed value to become corrupt and that points to somekind of error in the way the application loads checkpoint data.

Wow, good observation, this should help Travis find the bug and squish it :-)
4) Message boards : Number Crunching : Error with new longer tasks. (Message 4925)
Posted 27 Dec 2014 by Ananas
...
I was to fast thinking that restarting from a checkpoint is a problem.

I have just returned another result with a checkpoint restart and this one validated just fine....

I don't think that you're wrong, restarted (not resumed!) tasks are more likely to fail. It might depend on the moment when the checkpoint has been taken though.

Leaving suspended tasks in memory is a good idea here.

Depending on the projects you run concurrently, it might help to reduce the checkpoint frequency (write to disk every ...), especially projects with extreme HDD activity on checkpoints (MalariaControl is the worst one I know, but there might be others) might cause heartbeat errors, when multiple results checkpoint at the same time. If a checkpoint is used, you loose some more time, but the risk, that a checkpoint will be needed to resume work decreases.
5) Message boards : News : subproject page move completed (Message 4910)
Posted 18 Dec 2014 by Ananas
...Gotcha, I could make a redirect for that.

Before you go for each individual page ... a catch-all catch-most would probably be http://volunteer.cs.und.edu/subset_sum/.htaccess using the Apache Rewrite Engine, something like this should work :

RewriteEngine On RewriteBase /subset_sum RewriteRule ^(.*)$ /csg/$1
6) Message boards : Number Crunching : No finished file (Message 4887)
Posted 13 Dec 2014 by Ananas
There is an overwiew over the result series here
7) Message boards : News : [dna] new run working 'snail_hg19_1000fa_1motifs_3' (Message 4884)
Posted 13 Dec 2014 by Ananas
I pulled one ahead for a test by pausing some older ones :

gibbs_snail_hg19_1000fa_1motifs_3_372_0_0 ran without causing trouble, so this series looks more promising (Win XP x64 on Xeon)
The runtime matched the estimated time, result output has been created :
gibbs_snail_hg19_1000fa_1motifs_3_372_0_0_0 19.484 bytes gibbs_snail_hg19_1000fa_1motifs_3_372_0_0_1 4.104.036 bytes Outcome : Success Validate state : Valid
_____

You don't have to do the first test online, It is much easier (and less risky) to test at least one of each new series in standalone mode, so you can see strange output or other weird behaviour.

BOINC project apps usually allow this. Standalone mode is set, when no "init_data.xml" file is present in the working directory.
8) Message boards : News : [dna] new run up (Message 4883)
Posted 12 Dec 2014 by Ananas
gibbs_snail_hg19_1000fa_1motifs_2_232_0_2 shows Server state :In progress

instead of cancelled.
9) Message boards : News : [dna] new runs workunits/results removed (Message 4867)
Posted 12 Dec 2014 by Ananas
I had to stop BOINC and edit client_state.xml in order to suspend this project.

Only after that I regained control over the BOINC client, so I was able to remove all workunit containing the word "motifs".

It would have been better to cancel instead of just delete them, so the core clients would have known that they should kill the results.

@all, who have the same trouble ...try this (after exiting BOINC) :

<project> <master_url>http://volunteer.cs.und.edu/csg/</master_url> <project_name>Citizen Science Grid</project_name> ... <send_job_log>0</send_job_log> <suspended_via_gui/> <=== INSERT THIS LINE ... </project>


After a restart you will be able to maintain your result list and cancel the evil results
10) Message boards : Image and Video Discussion : Video watching ... fullscreen? (Message 4818)
Posted 10 Nov 2014 by Ananas
p.p.s.: I had one Tern video where the parents exchanged but there's no event for that (so I used "on nest" for the whole time). I also had one where a Grouse started to sort her "stuff" (while sitting on the nest), rearranging the grass - no event for that either. When the Grouse hen leaves the nest, I'm using "off nest" - would that have to be followed by a "not on video" after it has gone away?

The "events" box should be somewhat more talkative and have a brief explanation describing the events in simple words below it (there is enough unused space), so you don't have to stop watching in order to look for FAQs or dictionary.

Oh ... and the exact meaning of "foraging", would we even see if the bird was looking for food somwehere else? Wouldn't it simply be "off nest" and/or "not on video" then? (English isn't my 1st language so I might be missing some detail)
11) Message boards : Image and Video Discussion : Video watching ... fullscreen? (Message 4817)
Posted 10 Nov 2014 by Ananas
I just started last night and for me it seems to be the best thing, to set the speed to 5x, then switch to fullscreen and lean back.

When I think that something has changed, I pause the video and scroll back and forth across some timespan covering the event, until I'm sure what it was that caught my attention.

In my case it was a fly, a small bug and changing light and wind conditions (very often vegetation), not really notable events, but I think it is much easier to identify a relevant change at 5x speed than very smooth changings at 1x speed.

Switching back to 5x speed seems not to be supported in fullscreen (the choices I have there are 1x, 1.5x and 2x from the context menu), so then I have to ESC fullscreen mode, change the speed and re-enter fullscreen.

Is this method OK or is there a better one?

p.s.: For the video on the page, it would be nice to have the option to reduce the brightness of the page background color. The contrast between the (nightly) videos and the white background is very hard.
12) Message boards : Cafe : Suicide birds (Fasan / pheasant) (Message 4815)
Posted 10 Nov 2014 by Ananas
Pheasants are strange animals. There is a place in Sweden called Trolle Ljungby where people breed them and the pheasants keep running across the street without caring for anything. The result is this typical situation :-(

Not my images but I have been there a few times and even when you drive along the road in 1st gear as slow as possible, those guys just walk in front of your car, so you have to stop and wait. I even exited the car in order to chase them away but that didn't impress them too much either.

Most drivers (especially trucks) do not slow down at all and just drive over them, the road is full of dead pheasants all the time. There are (funny) warning signs but it is a busy road and people who are not on vacations seem to be always in a hurry there.

There are many ducks too there but those are way more careful, it's usually just the pheasants that end up as roadkill.

Very strange behaviour for birds - but also for the guys who breed them there :-/
13) Message boards : News : new runs started up (Message 4814)
Posted 9 Nov 2014 by Ananas
No trouble with any of them (I paused some in order to try a full mix)

Edit ... maybe worth to mention : Slug 1000 run not even half as long as Test 1000 and Snail 1000 (but of course slugs don't have to carry houses so they can run faster)
14) Message boards : Webpage and Application Code Discussion : Application page (Message 4813)
Posted 9 Nov 2014 by Ananas
Imo. the "applications" page should be moved to the "Information" pulldown next to "Server Status". It is not user related.
15) Message boards : Number Crunching : Heartbeat errors destroy results (rare problem) (Message 4812)
Posted 9 Nov 2014 by Ananas
You're right, the checkpoints are supposed to catch and fix the heartbeat problems.

If you need one more sample for checking : hostid=4451

edit : Here's an invalid one with a regular restart (not heartbeat related) : resultid=630168


@all : If possible, use this setting :

Leave tasks in memory while suspended? yes

It reduces the risk for invalid results.
16) Message boards : Number Crunching : Heartbeat errors destroy results (rare problem) (Message 4799)
Posted 5 Nov 2014 by Ananas
I just found a wingman who has BOINC heatbeat trouble :

hostid=4986

Currently he has 2 invalid, 2 error and one inconclusive, at leat 4 of them show restarts caused by heartbeat problems.

From what I can see, this seems to be a rare problem, so it probably doesn't require immediate investigation.


p.s.: NOT related to the problem with the strange output like this:<stderr_txt> 0,0,0,0,0,0,0,0,0,
17) Message boards : Number Crunching : Marked invalid, I've got a question. (Message 4742)
Posted 24 Oct 2014 by Ananas
This is probably really bad :

http://volunteer.cs.und.edu/csg/workunit.php?wuid=182132

In this case, two of those strange results ended up in the scientific database and the one with the normal looking stdout became invalid.
18) Message boards : Number Crunching : Marked invalid, I've got a question. (Message 4738)
Posted 21 Oct 2014 by Ananas
Not only Linux boxes are affected :

Same error type on Windows 8.1 x64 (not my machine), that means that it cannot be an issue with the 0.47 built (there is no 0.47 on windows).

Looks much like an uninitialized variable or an array that is filled across it's limits, messing up a variable that is behind it in memory.

All machines have mostly normal looking results and bail out only seldom.

p.s.: Here would be one more, Windows 7 x64 this time

And this box has several on Linux x64, all 0.47 this time.
19) Message boards : Number Crunching : How accurate is the server status page? (Message 4681)
Posted 15 Oct 2014 by Ananas
Interesting, 4,738 new users in the last 24 hours but only 12 computers added also in 24 hours.

Possibly new Spam accounts? (With a few actual users added to the mix).

Conan

In theory they could have been created by activating the CSG accounts of subproject users (which then would not have hosts attached yet), but I checked some in the current high numbers, the names are a random mix of upper-/lowercase letters and have no equivalent on BOINCstats.

There are still lots of subproject users who haven't activated their CSG account yet.
20) Message boards : News : anyone else missing credits after the swap to the new boinc per-app credit system? (Message 4673)
Posted 8 Oct 2014 by Ananas
Dirk didn't refer to the handling but to the latest refresh, that's why he mentioned the last update timestamp - it's more than a week old


Next 20