Project Update

[JugNut]

It's all good kashi as it should be the project admins job to inform people of problems,  you know in the notices section in Boinc manager or the news thread but nothing!

I read the "extremely long running tasks" thread before I posted there but there was nothing about about constantly resetting tasks just really long running ones,  at least thats I could find, which is why I started another thread. 

start rant...
How long has Krystof known about the useless v0.02 tasks?  If he has known of this for many hours then why did he say nothing?  Heck why hasn't he canceled the whole lot of them?  Last I checked there still being sent out now.
In his reply to my post he finally tells me to abort everything,  it would have been nice if he had mentioned that hours earlier and stopped all users(and me) from wasting their hard earned time & money
end rant....

Ah that feels better 
OK that was a poor rant i'll try harder next time :)

But your right kashi I assumed your post was only about the long running task problem. 

So... WCG here I come..

[kashi]

I thought your rant was quite good actually, not poor at all.    

Yes I got caught too, although he mentioned the bug in Version 0.2 tasks at 11 am yesterday, I didn't realise it affected all 0.02 tasks so only aborted any that hadn't been started. The 17 that had already started I left running alongside some UV tasks. So wasted those crunching resources until I caught them after about 6 hours and aborted them.

I knew all 0.02 tasks were buggy then, not just a few, because I checked my wingmen and not one had returned a valid task even the previously speedy single core Anonymous hosts.  Wasn't too upset as other cores had still been crunching UV tasks so all was not lost. Lot more work/electricity lost for you though, your rant was thoroughly justified!

When I switched from Linux to Windows and immediately received an 0.02 task, I knew something was badly awry and the buggy 0.02 tasks were still on the server getting handed out. Was disappointed that a more prominent warning to abort them was not given by admin in News or on forum.  Decided not to post about it myself in Universe forum then because my sneaking suspicion made a reappearance because of the timing. Thought it was possibly distraction tactics: "Look over there, buggy tasks!"

That's when I pulled the plug on Universe and went elsewhere. Hadn't mentioned this latest suspicion before because it's most uncharitable of me to ascribe to malice what is most likely just incompetence, which is unfortunately all too common on BOINC projects. I mean there's still 26,421 0.02 tasks on the server, don't know how many of the 73,447 tasks in progress are 0.02 versions but considering all those people running with a smaller cache it could possibly be a high percentage. That's a terrible lot of wasted resources and electricity.

WCG eh, you can add a nice Smash Childhood Cancer hammer badge to your excellent collection.  
 

 

[Dataman]

They must not send these bad wu's to we lowly RPi crunchers. I have had no problems at all other than one I aborted because it ran 25+ hours and was only 50% complete. Maybe you blokes should order up some Raspberry Pi (would you like ice cream and whipped cream on that?)
 

[chooka03]

Ah good to see some of the big guns out crunching U@H 
I've been trying to get the team up the ranks for some time as U@H is a laggard for the team. That dam team Italy was juuuuuuussstttt keeping me at bay but now we'll slam them and overtake the other quickly.

The guys at U@H have about 2Million.....yes 2 Million WU's now! It's awesome to finally get some stable WU's.
I'm really looking forward to the team gaining some serious spots.. (I've added the team link below for quick reference)

Crunch on brothers!!
 

[JugNut]

Did anyone else have problems returning BHDB tasks yesterday? My uploads were going so slow that my PC's  started running out of work.  When your uploads get stuck or slow down enough and reach a certain number the server no longer allows you to download. So even though there were millions of WU's available plus my PC's were running dry too,  I still received bupkis. But not for long...

So instead of letting my PC's sit and do nothing I decided to try ye ol' back door approach and fired up the extra boinc instances that I use mainly for moneys NCI work and attach to universe using one of those on each affected PC,  which worked well and got them back to work quick smart.

Actually now that I think about it the above method could have backfired badly "if" there would have been a huge supply of Black hole WU's, as then the same thing would have just keep repeating over & over and even got worse.  But instead I received BHspin2 tasks which take longer to do and have a much smaller upload file which ended up giving the BHDB tasks time to finally upload.  Also as time went on & less & less BHDB task where being download and returned to the server, the server started speeding up.
So eventually I got them all to upload some time early this morning. Today however the server speeds are showing a vast improvment on yesterday.

Although they might be slow to upload those BHDB tasks sure do credit nicely ;)

Needless to say when much of yesterday's work eventually uploaded and then combined with today's lot it's starting to look like a very nice credit day.

Crunch-on..

[chooka03]

I had trouble uploading WU's for a while there Jugnut..... I used this method which worked.

https://universeathome.pl/universe/forum_thread.php?id=269&sort_style=&start=20

[JugNut]

Yep you're right chooka that option saved my bacon once before,  i'm sure I posted about it recently? 
Anyhoo, as far as I know it's already activated on all my hosts.



EDIT:
Oh that's right.. (Second post from the bottom)
http://nci.goofyxgridathome.net/forum_thread.php?id=284&postid=2544#2544

At the time of the above post the difference between hosts that either had the option set in cc_config and those that did not was like night & day. Those that had it worked fine those that did not got nothing but errors.

[kashi]

Haha, yep those BHDBs certainly credited like crazy. Nice big credit dump when the server came back up recently and could finally upload all those completed tasks.

Had switched elsewhere when BHDBs ran out but saw chooka's call to arms and aborted Asteroids cache and switched back to Universe. Had to tick BHspin application in preferences, "accept work from other applications" was not sufficient to get any. BHspin tasks are still reasonably generous even if not up at BHDB level.

Will put dual box on half Universe, half Sourcefinder when they release a new Sofia batch. Would like to get it to 500K before it finishes even though it is a horror, with VMs, huge memory requirements and a terrible wonky validator resulting in far too many errors.

Ah, the old http proxy setting trick, remember having to set this occasionally many years ago on some projects with my TPG ADSL connection, even though TPG repeatedly denied ever using a transparent proxy.

[chooka03]

We should really fly up the ranks now! Already up 5 places this month 

Go team!

 

[JugNut]

Yea if we can keep up something close this pace we should be moving up soon enough 

@ Kashi: Yea i'm still on ADSL2 with Telstra.  I've had an awful lot  of issuses over the past 6 months, just when I think they've got all the bugs finally sorted out a new one appears. It's never ending.

Even though i'm in the middle of sydney the NBN seems to be in all suburbs except ours. On the NBN rollout map our suburb looks like a NBN free island? For some reason it's not even slated to start for another 2 years.

By the time it finally gets installed 5G will out and if the pricing & speed is right I may not even need NBN.


PS: Hmm come to think of it perhaps TPG was leasing from telstra back then? LOL

[tazzduke]

Woohoo, its like a miny AA lol, just threw 2 of my pcs onto Universe, another project I have not done any for a long time lol

[kashi]

Haha, yay, mini aa, rah, rah, rah.

@JugNut Well with nbn™, sometimes delay can be better in the long run, even if it doesn't seem so at the time. Don't want to get your hopes up unnecessarily but although your address shows as planned for FTTN on nbn™ maps, Mr Mac's map shows you as FTTC (FTTdP). As you are probably aware, FTTC although not FTTP is still much better than being stuck on FTTN, which suffers greatly in many cases from poor speed and unreliability due to using long runs of old copper wiring.

Plus FTTN is not upgradeable and must be replaced, whereas an FTTC upgrade to either G.fast or FTTP is relatively easier. It's even possible the newer G.fast capable FTTC DPUs (in your pit) may be being installed by the time they finally get to your street. If so, in theory G.fast speeds would be available immediately. The NCD (modem/reverse power box in house) is already G.fast capable apparently.

The nbn™ maps are often out of date and were reportedly due for an update on 18th April, but it hasn't happened so far today, so currently nobody knows when they will update. Yes, just updated now, see edit below.

There was absolutely no valid technical or business case for using the end of life copper telephony network in Australia as the basis for a national data network in 2013. Although reported rarely, if ever, it is well known by anyone with any unbiased networking knowledge that FTTN was only employed in Australia due to corruption.

Edit: nbn™ map has just updated, as predicted by Mr Mac's map your address has changed from FTTN to FTTC. Still a few more years to wait unfortunately:

Planned | nbn™ Fibre to the curb (FTTC)
Planned availability: Jan-Jun 2020*
The rollout of the nbn™ is being planned in this area.

[JugNut]

That's some awesome detective work there kashi, thanks for the effort you put in it's very much appreciated. 
And your right being left behind may actually prove to be a massive benefit in the long term.

Thanks you so much, you made my day :)    + karma