Progress?

Message boards : Number crunching : Progress?
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
itenginerd
Avatar

Send message
Joined: 1 Aug 00
Posts: 37
Credit: 39,905
RAC: 0
United States
Message 211634 - Posted: 12 Dec 2005, 16:45:46 UTC

In what I remember seeing from late last week, I haven't seen the validation queue grow much at all... I would expect that we would have broken 400,000 over the weekend if it was just a function of some uploads coming in to ease the load. Granted, I could be misremembering--somebody let me know if you've got real numbers on this!

1) Are the stats pages not updating correctly? They've worked up to now, so I doubt that they're suddenly out to lunch. (Got some graphs Scarecrow?)

2) Are we actually successfully DoS'ing the UL server? What are the options to fix that, S@H staff?

Where do we go from here--because it's somewhat apparent that we aren't going anywhere right now with the status quo... 366,908 results in VQ as of 16:40 UTC 12/12/05.

(j)
James



ID: 211634 · Report as offensive
Profile ML1
Volunteer moderator
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 20142
Credit: 7,508,002
RAC: 20
United Kingdom
Message 211655 - Posted: 12 Dec 2005, 17:15:32 UTC - in response to Message 211634.  
Last modified: 12 Dec 2005, 17:17:43 UTC

...1) Are the stats pages not updating correctly? They've worked up to now, so I doubt that they're suddenly out to lunch. (Got some graphs Scarecrow?)

Yes, most of the numbers are out of date. Various back end services are disabled or otherwise not functioning while the present glitches are cleared.

This also means that all current WU results are not going to time out regardless of their deadlines.

2) Are we actually successfully DoS'ing the UL server? What are the options to fix that, S@H staff?

Yes, although the Boinc design is meant to avoid this. I think some assumptions on the exponential backoff numbers that is meant to avoid this are a little off... Hang in there for the next update!

Where do we go from here--because it's somewhat apparent that we aren't going anywhere right now with the status quo... 366,908 results in VQ as of 16:40 UTC 12/12/05.

That depends on what you can see or not.

The data upload rate is increasing although it is still very erratic.

Berkeley will fathom where the latest bottleneck is and produce a fix. They'll likely also add further improvements to make the Boinc system more resiliant for this type of problem.

Meanwhile, we just need to chill out and watch the story unfold. There's other Boinc projects to take up the slack in the mean time.

Happy crunchin',
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 211655 · Report as offensive
Idefix
Volunteer tester

Send message
Joined: 7 Sep 99
Posts: 154
Credit: 482,193
RAC: 0
Germany
Message 211796 - Posted: 12 Dec 2005, 19:25:05 UTC - in response to Message 211634.  

Hi,
Where do we go from here--because it's somewhat apparent that we aren't going anywhere right now with the status quo... 366,908 results in VQ as of 16:40 UTC 12/12/05.
The scheduler is not responding --> These graphs are meaningless in the moment.

Regards,
Carsten
ID: 211796 · Report as offensive
Ulrich Metzner
Volunteer tester
Avatar

Send message
Joined: 3 Jul 02
Posts: 1256
Credit: 13,565,513
RAC: 13
Germany
Message 211820 - Posted: 12 Dec 2005, 19:58:40 UTC

Ok, and what happens to the results i have to leave on the computer @ work, when i go into holidays on next thursday and will not return before the next year?

Thanks Berkeley, for once again wasting my time and money... :/
Aloha, Uli

ID: 211820 · Report as offensive
Profile Borgholio
Avatar

Send message
Joined: 2 Aug 99
Posts: 654
Credit: 18,623,738
RAC: 45
United States
Message 211823 - Posted: 12 Dec 2005, 20:05:08 UTC - in response to Message 211820.  

Ok, and what happens to the results i have to leave on the computer @ work, when i go into holidays on next thursday and will not return before the next year?

Thanks Berkeley, for once again wasting my time and money... :/



Are you the kind of person who volunteers your spare time to the Red Cross and then get pissed because you're not getting paid?
You will be assimilated...bunghole!

ID: 211823 · Report as offensive
Ulrich Metzner
Volunteer tester
Avatar

Send message
Joined: 3 Jul 02
Posts: 1256
Credit: 13,565,513
RAC: 13
Germany
Message 211827 - Posted: 12 Dec 2005, 20:08:09 UTC - in response to Message 211823.  

Are you the kind of person who volunteers your spare time to the Red Cross and then get pissed because you're not getting paid?

That's exactly the kind of answer i expected from an expert like you...
Aloha, Uli

ID: 211827 · Report as offensive
Profile Borgholio
Avatar

Send message
Joined: 2 Aug 99
Posts: 654
Credit: 18,623,738
RAC: 45
United States
Message 211833 - Posted: 12 Dec 2005, 20:19:29 UTC - in response to Message 211827.  

Are you the kind of person who volunteers your spare time to the Red Cross and then get pissed because you're not getting paid?

That's exactly the kind of answer i expected from an expert like you...



It doesn't take an expert to know what "volunteer" means. It means you donate your spare time / money / effort to a cause, without expecting any payment in return. Berkeley never forced you to donate anything. You did it of your own free will. So if you consider this a waste, it's your own dumb-a$$ fault for donating. Not Berkeley's.
You will be assimilated...bunghole!

ID: 211833 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 211837 - Posted: 12 Dec 2005, 20:23:30 UTC - in response to Message 211820.  
Last modified: 12 Dec 2005, 20:30:24 UTC

Isn't volunteering the same in Germany? How come an unforseen problem at Seti made you crunch extra for electricity and time?

What time? The otherwise idle time? Or the otherwise time the computer would have been off? Which you couldn't have done by yourself, because someone from Berkeley pressured you into signing a contract that you had to crunch Seti units for all this time... am I correct?

I wonder who you blame when the bus you use to go to the airport will break down, due to which you will miss your airplane.

Will you blame yourself for taking the bus?
Will you blame the bus driver for not continuing?
Will you blame the bus company for not replacing the bus?
Will you blame the company you fly with for not halting the plane until you got on (3 days late)?
Will you blame the airport for not keeping your plane on the tarmac until you were on?


i have to leave on the computer @ work ...

...once again wasting my time and money...

Now I do wonder, how do these two relate? Is your boss demanding money from you for using the computer at work? Or are you boss?
ID: 211837 · Report as offensive
Ulrich Metzner
Volunteer tester
Avatar

Send message
Joined: 3 Jul 02
Posts: 1256
Credit: 13,565,513
RAC: 13
Germany
Message 211844 - Posted: 12 Dec 2005, 20:30:35 UTC

unforseen? hahaha!

Spare that cheerleading crap for noobs, i already know it by heart...

bye, happy Xmas'n stuff
Aloha, Uli

ID: 211844 · Report as offensive
Profile Borgholio
Avatar

Send message
Joined: 2 Aug 99
Posts: 654
Credit: 18,623,738
RAC: 45
United States
Message 211862 - Posted: 12 Dec 2005, 20:44:26 UTC - in response to Message 211844.  

unforseen? hahaha!

Spare that cheerleading crap for noobs, i already know it by heart...

bye, happy Xmas'n stuff



Cheerleading? You mean they should have the ability to forecast server problems well in advance? You're saying they should have the magic talent to predict every crash / breakdown / power surge before they happen? You really ARE stupid. Merry Xmas? Sheesh...Santa won't even bother bring you coal. You'd be lucky if you got run over by his sleigh.
You will be assimilated...bunghole!

ID: 211862 · Report as offensive
Whiskey
Volunteer tester
Avatar

Send message
Joined: 5 Sep 04
Posts: 981
Credit: 640,589
RAC: 0
United Arab Emirates
Message 211863 - Posted: 12 Dec 2005, 20:45:45 UTC

Hey,
Here's a great idea.
Why don't you just leave BOINC running, turn off your Monitor for a few days, get out of the house, go to the Pub, Sports Centre or do whatever you do when you are'nt staring at the Monitor and the TV.
There is a life outside your house, there are people out there who have never heard of SETI, BOINC and Distributed Computing and don't give a damn about it.



Join the #1 UAE Team.
ID: 211863 · Report as offensive
Zap de Ridder
Volunteer tester

Send message
Joined: 9 Jan 00
Posts: 227
Credit: 1,468,844
RAC: 1
Netherlands
Message 211872 - Posted: 12 Dec 2005, 21:03:02 UTC

One thing I can tell, wus getting reported right now.
ID: 211872 · Report as offensive
harald hermann

Send message
Joined: 21 Jun 99
Posts: 10
Credit: 170,618
RAC: 0
Germany
Message 211882 - Posted: 12 Dec 2005, 21:09:44 UTC - in response to Message 211872.  

One thing I can tell, wus getting reported right now.


yes you are right.
only one to upload
reports send
but still no download

ID: 211882 · Report as offensive
Daniel Schaalma
Volunteer tester
Avatar

Send message
Joined: 28 May 99
Posts: 297
Credit: 16,953,703
RAC: 0
United States
Message 211897 - Posted: 12 Dec 2005, 21:25:21 UTC - in response to Message 211862.  

Cheerleading? You mean they should have the ability to forecast server problems well in advance? You're saying they should have the magic talent to predict every crash / breakdown / power surge before they happen? You really ARE stupid. Merry Xmas? Sheesh...Santa won't even bother bring you coal. You'd be lucky if you got run over by his sleigh.


I guess Matt L. lost his Carnak the Magnificent hat. Can't work magic and see into the future without one...

Regards, Daniel.
ID: 211897 · Report as offensive
Profile Borgholio
Avatar

Send message
Joined: 2 Aug 99
Posts: 654
Credit: 18,623,738
RAC: 45
United States
Message 211959 - Posted: 12 Dec 2005, 22:02:31 UTC - in response to Message 211897.  

Cheerleading? You mean they should have the ability to forecast server problems well in advance? You're saying they should have the magic talent to predict every crash / breakdown / power surge before they happen? You really ARE stupid. Merry Xmas? Sheesh...Santa won't even bother bring you coal. You'd be lucky if you got run over by his sleigh.


I guess Matt L. lost his Carnak the Magnificent hat. Can't work magic and see into the future without one...

Regards, Daniel.



Maybe the complainers can pitch in and get him one for an early Xmas present.
You will be assimilated...bunghole!

ID: 211959 · Report as offensive
itenginerd
Avatar

Send message
Joined: 1 Aug 00
Posts: 37
Credit: 39,905
RAC: 0
United States
Message 212158 - Posted: 13 Dec 2005, 0:37:40 UTC - in response to Message 211655.  
Last modified: 13 Dec 2005, 0:41:58 UTC

<sigh> I really didn't mean my original post as flame bait. It was meant as a couple of legitimate questions. I've got E@H, which has been rock stable lately, so the current S@H problems are not a big worry. Commentary follows.

@ML1: I find your reply unhelpful, and a touch condescending. I don't know if you meant it that way, but that's how I read it. Unless you think I'm a complete newbie, what you said doesn't get us any closer to a solution or in any way answer the original questions I posed.

Yes, most of the numbers are out of date. Various back end services are disabled or otherwise not functioning while the present glitches are cleared.

This also means that all current WU results are not going to time out regardless of their deadlines.

Just because the validators are off doesn't mean we can't count what's in the validator queue. Not only can I read the server status page, but I can also read the front page news, the technical news, and lots and lots of whiny, complaining, or argumentative posts in the fora on the matter. Apologies if my post came across as just one more of them--wasn't my intention. My point, which you either missed or ignored, was that if the validator queue figure is right (which it is until I hear a reason why it's wrong), it isn't increasing very quickly. Until it increases significantly, we won't be out of the woods on this outage. Hence the reference to Scarecrow's graphs from the last big outage.

Yes, although the Boinc design is meant to avoid this. I think some assumptions on the exponential backoff numbers that is meant to avoid this are a little off... Hang in there for the next update!

If we're waiting on a new BOINC CC release to fix this problem we're in deeeep $#!+. Unless you meant the next news update, which we will assuredly wait for. I doubt we'll hear anything useful till the Wed database downtime, though--just a hunch.

The data upload rate is increasing although it is still very erratic.

Let's not confuse DATA rate with UPLOAD rate. If I get half a result uploaded before the connection times out, then my traffic is raising the DATA rate (which is what Cricket measures) without changing the UPLOAD rate (which is only measured by the validator queue, afaik). One of the things I have noticed is that the first packet of many of my uploads is getting through before the connection times out. Not only does this keep a socket on the UL server open longer, it contributes useless traffic to the Cogent link.

Meanwhile, we just need to chill out and watch the story unfold. There's other Boinc projects to take up the slack in the mean time.

I wholeheartedly agree. Making whine was not my intention.

@Idefix: I completely disagree with you for the reasons mentioned above. Just because the bottle is corked doesn't mean we don't know how much grog is in it.

@Uli: While I can sympathize, just remember: we're competing for imaginary numbers. If you think Berkeley's bad, I'd hate to hear you rant about governement... 8)

Frankly, nothing else in this thread is worth commenting on.

VQ: 455,341 as of 00:20 UTC

(j)
James
ID: 212158 · Report as offensive
forward

Send message
Joined: 1 Jun 99
Posts: 53
Credit: 2,658,479
RAC: 0
United States
Message 212167 - Posted: 13 Dec 2005, 0:43:02 UTC - in response to Message 211823.  
Last modified: 13 Dec 2005, 0:45:18 UTC

Ok, and what happens to the results i have to leave on the computer @ work, when i go into holidays on next thursday and will not return before the next year?

Thanks Berkeley, for once again wasting my time and money... :/



Are you the kind of person who volunteers your spare time to the Red Cross and then get pissed because you're not getting paid?

If you give money to the Red Cross and it loses it in a poker game, it is the same as Berkely throwing away the time you have spent working on Seti.
ID: 212167 · Report as offensive
Ingleside
Volunteer developer

Send message
Joined: 4 Feb 03
Posts: 1546
Credit: 15,832,022
RAC: 13
Norway
Message 212174 - Posted: 13 Dec 2005, 0:48:57 UTC - in response to Message 212158.  
Last modified: 13 Dec 2005, 0:54:00 UTC

My point, which you either missed or ignored, was that if the validator queue figure is right (which it is until I hear a reason why it's wrong), it isn't increasing very quickly. Until it increases significantly, we won't be out of the woods on this outage. Hence the reference to Scarecrow's graphs from the last big outage.



As also shown on the Status-page, the Transitioners is currently over 3 hours backlogged. Since it's the Transitioners job to set the NEED_VALIDATE-flag, there can currently be many reported results that still isn't added to the Validator-queue.

Also, since Validator-queue is a wu-count and not result-count, it will not increase as fast "In progress" is dropping. With many backlogged result-uploads, there can be many wu with only 1 or 2 results, meaning not ready for validation, and therefore not showing-up in Validator-queue. Others is 4th result for wu already in Validator-queue.


Anyway, just had a result-upload claiming speed 25KB, so it looks like things have improved. :)

ID: 212174 · Report as offensive
Profile Nebby
Avatar

Send message
Joined: 14 May 99
Posts: 54
Credit: 1,896,156
RAC: 0
United States
Message 212180 - Posted: 13 Dec 2005, 0:53:26 UTC

I have seen quite a few of my results upload recently. I think we are are track for a recovery. In fact, it even looks like things are improving rapidly on my end.
ID: 212180 · Report as offensive
Profile ML1
Volunteer moderator
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 20142
Credit: 7,508,002
RAC: 20
United Kingdom
Message 212208 - Posted: 13 Dec 2005, 1:08:39 UTC - in response to Message 212158.  
Last modified: 13 Dec 2005, 1:13:27 UTC

<sigh> I really didn't mean my original post as flame bait...

Good and noted.

@ML1: I find your reply unhelpful, ... Unless you think I'm a complete newbie, what you said doesn't get us any closer to a solution or in any way answer the original questions I posed.

As far as I saw, I tried answering in kind. Sorry, no condescention was intended.

Yes, most of the numbers are out of date. Various back end services are disabled or otherwise not functioning while the present glitches are cleared.

This also means that all current WU results are not going to time out regardless of their deadlines.

Just because the validators are off doesn't mean we can't count what's in the validator queue....

There's enough of the back end services switched off or disrupted that the numbers visible to us cannot be relied upon, as noted here in this same thread.

...Until it increases significantly, we won't be out of the woods on this outage. Hence the reference to Scarecrow's graphs from the last big outage.

That requires a few assumptions or a good understanding of what's happening. More than I know from here.

If we're waiting on a new BOINC CC release to fix this problem we're in deeeep $#!+.

Not at all.

The server side stuff gets fixed to solve the immediate problem(s). Learning from the past few days, the Boinc client then gets updated some time later so that the DDOS type effects are mitigated more gracefully or some other fix is put in place.

The data upload rate is increasing although it is still very erratic.

Let's not confuse DATA rate with UPLOAD rate...

Exactly, hence the words I used.

Aside, there's just been a big jump up to 45Mbit/s just now. Somethings changed...

Meanwhile, we just need to chill out and watch the story unfold. There's other Boinc projects to take up the slack in the mean time.

I wholeheartedly agree. Making whine was not my intention.


Good comment to end on!


Happy crunchin',
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 212208 · Report as offensive
1 · 2 · 3 · Next

Message boards : Number crunching : Progress?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.