Looks like Seti is down again since last Friday. I can't believe it. These guy's at Seti - Boinc have got to be the most unlucky bunch of people that was ever assembled into one group. Nothing, I mean nothing, goes easy for them. Please don’t get the wrong idea here. I think they have done a wonderful job so far with Boinc and Seti. It’s just that if it wasn’t for all the bad luck with the hardware, they wouldn’t have any luck at all.

Let’s use this thread to give the guy’s some encouragement and some praise for their work and cross your fingers that their luck will change soon!

Don

---

Boinc....Boinc....Boinc....Boinc....

> Let’s use this thread to give the guy’s some encouragement and some praise for
> their work and cross your fingers that their luck will change soon!
>
> Don

Terrific idea! I've already stressed in another posting that the question "Is anybody out there?" becomes more and more important, and this is a sure sign of our progress as a civilization. The S@H team is doing work of great significance. If we all stuck only to projects of immediate practical usefulness, where would we be now? Somewhere about the level of the Roman empire, max.
I admire their will to go on no matter what obstacles are on their way. Sorry to have joined the project so late - up to several months ago I had a very obsolete PC that would not make any difference. But they can count on me in the future, and I'm already thinking about a serious upgrade in order to help a little more.
Have fun crunching!
Vladimir

I hope you are not misled by this "1999" under my name. Don't know how it appeared there.

---

I agree. This project has just about all the bad luck one can imagine.
None of this is down to design or implementation issues, just bad luck and presumably budget constraints.

It has to be remembered that the original SETI had nothing like the issues faced by BOINC when it comes to load. At start-up, seti 1 had a few thousands of users, not the 10s/100s of thousands SETI/Boinc has. Worse, remember at SETI 1 start, the average user had what - a 133Mhg Pentium and screen-saver software - crunching 1 WU in days or even weeks. Now, PCs require to be fed between 4 and 20+ WUs a day.

Well done to the BOINC team - I'm sure it'll settle in soon.

---

@ Vladimir:
Did you run SETI classic? If so, then that's why the Joined: date is Apr. 3, 1999. That was the date you signed up with the original SETI. If you didn't run Classic, I haven't a clue.

@ Geek@Play:
It's nice to see some positive statements for a change. I have spent 2 days trying to convince these nay-sayers that this current problem isn't SETIs. They lost their Internet connection through Cogent. The whiners just wouldn't listen to me or others.

@ Whomever:
I have been with SETI since May 23, 1999. I switched to BOINC/SETI last July and haven't looked back. Classic SETI will be closed down in about a month or so, from what I have read. I'm in this for the science, not the credits or the competition. I stand behind SETI and want to help them achieve their goal, finding E.T. I'm willing to trudge through the ups and downs. Knowing that BOINC is relatively young, it will have growing pains. Over the past few months, it has gotten better. I give the Dev Team and Staff, 2 BIG thumbs up for sticking with it through thick and thin. I can do no less....

L8R....

T'Khasi Time: Monday, 21 March 2005 - 06:15 PM --800 (Pacific Standard Time)

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

> I agree. This project has just about all the bad luck one can imagine.
> None of this is down to design or implementation issues, just bad luck and
> presumably budget constraints.

If anything, it validates the basic BOINC structure because:

1) Those who set "connect every 'x' days" to something high enough are likely still crunching.

2) Those who are crunching multiple projects have kept crunching all the way through, just a different "mix" of work.

... and the ideas behind BOINC seem to be working very well.

Note too that the circuit giving us fits right now is the Cogent circuit installed for SETI@Home in 2002. So it's actually a "classic" facility that's down.

---

> @ Vladimir:
> Did you run SETI classic? If so, then that's why the Joined: date is
> Apr. 3, 1999. That was the date you signed up with the original SETI. If you
> didn't run Classic, I haven't a clue.

I crunched less than 200 Classic WUs since last October - under a nick with a different e-mail address - then I switched to BOINC. So it's definitely not 1999.

---

The date thing: I do believe, since I did run classic and my date has been corrected, that the april 3rd 1999 date are the ones who have not run classic - or at least hasn't got a "link" between their boinc and classic accounts.

About the outage: this weekend there actually weren't many projects to crunch for. Seti was cut off, predictor was shut down, lhc had no work, pirates had a little which isn't out of the ordinary, burp isn't that far yet and actually shut their scheduler down "for maintanance" which left einstein and climate as the only ones to crunch for. Ned, you're absolutely right, BOINCs ability to crunch for several projects has proven it's worth - but this weekend has also proven that if you want to be absolutely sure your CPU doesn't end up at idle state, it's actually not enough just to run two projects.

Well, this weekend turned out to be a big inspiration for me personally as I found the extra motivation to make a scheduler status script for my website - a danish saying says "nothing is so bad that it isn't good for something" ;)

---

Actually CPDN was down for part of the weekend also, but hardly anyone noticed because of their humongous work units.

Absolutely agree, best to join as many projects as your machine(s) can handle within deadlines. Besides avoiding idle time it's also more fun. Each project is interesting in completely different ways; boredom is impossible. (At least for me, both as a software geek and one who enjoys science in general.)
And when things go wrong, you can watch BOINC ride that out as it's designed to do.

Kudos to all the platform and project developers, site maintainers, documenters, forum-helpers etc.

---

WOW, I just a new Seti work unit. Its starting to work, YEAH !!!

---

BOINC SYNERGY is an International Team and We Welcome All BOINC Participants!
BOINC Synergy Click to Join BOINC Synergy

> WOW, I just a new Seti work unit. Its starting to work, YEAH !!!
>
>
--- - 2005-03-21 21:36:32 - May run out of work in 5.00 days; requesting more
SETI@home - 2005-03-21 21:36:32 - Requesting 641369 seconds of work
SETI@home - 2005-03-21 21:36:32 - Sending request to scheduler: http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
SETI@home - 2005-03-21 21:36:53 - Scheduler RPC to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed
SETI@home - 2005-03-21 21:36:53 - No schedulers responded
SETI@home - 2005-03-21 21:36:53 - Deferring communication with project for 14 minutes and 39 seconds

I don't think so

---

3/21/2005 10:28:58 PM|SETI@home|Finished download of 29dc04aa.13138.17346.479830.181
3/21/2005 10:28:58 PM|SETI@home|Throughput 8062 bytes/sec


I must be lucky then.

---

BOINC SYNERGY is an International Team and We Welcome All BOINC Participants!
BOINC Synergy Click to Join BOINC Synergy

> Ned, you're absolutely right, BOINCs ability to
> crunch for several projects has proven it's worth - but this weekend has also
> proven that if you want to be absolutely sure your CPU doesn't end up at idle
> state, it's actually not enough just to run two projects.

Well, on Thursday I was crunching two projects that were up (one down) and now I've got three that are up. It never went below 1.

(and why do people crunch "pirates" anyway?)

---

I can only imagine how unmercifully hard all our machines will be hammering them for many hours to come, after this extended outage.
So yeah you're lucky. Enjoy :)

>(and why do people crunch "pirates" anyway?)
'cause it's entertaining, and it helps Einstein/BOINC development

> http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed
> SETI@home - 2005-03-21 21:36:53 - No schedulers responded
> SETI@home - 2005-03-21 21:36:53 - Deferring communication with project for 14
> minutes and 39 seconds
>
> I don't think so

This is typical when SETI starts back up -- according to cricket they're running a bit over 85 megabits/sec., well above the typical 45-ish.

But I think you've seen this before....

---

> 3/21/2005 10:28:58 PM|SETI@home|Finished download of
> 29dc04aa.13138.17346.479830.181
> 3/21/2005 10:28:58 PM|SETI@home|Throughput 8062 bytes/sec
>
>
> I must be lucky then.
>
http://setiweb.ssl.berkeley.edu/forum_thread.php?id=12754

---

> > ....
> I crunched less than 200 Classic WUs since last October - under a nick with a
> different e-mail address - then I switched to BOINC. So it's definitely not
> 1999.
>

Well Vladimir, let's see if we can figure this out....

@ Ned Ludd:
Ned, I noticed this same thing in another thread were you were participating. Now, don't think I'm blaming you, I'm not! I forget which thread just now, but there were at least 3 users there that had Apr. 3, 1999 as their Joined: date. It's happening here too. Any ideas? This seems so weird. >:-/ Vladimir says that the date is wrong for him.

L8R....

T'Khasi Time: Monday, 21 March 2005 - 07:53 PM --800 (Pacific Standard Time)

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

> @ Ned Ludd:
> Ned, I noticed this same thing in another thread were you were participating.
> Now, don't think I'm blaming you, I'm not! I forget which thread just now,
> but there were at least 3 users there that had Apr. 3, 1999 as their
> Joined: date. It's happening here too. Any ideas? This seems so
> weird. >:-/ Vladimir says that the date is wrong for him.

I saw another post that said that the date changed on all BOINC accounts that were not "brought over" from Classic.

I started a new BOINC account because I have no idea what my E-Mail address might be on Classic. It probably isn't even valid.

---

> > ....
> I saw another post that said that the date changed on all BOINC accounts that
> were not "brought over" from Classic.
>
> I started a new BOINC account because I have no idea what my E-Mail address
> might be on Classic. It probably isn't even valid.
>

If I understand correctly, it's only happening to users that create a new BOINC account vs. activating their existing account from Classic. Right? Still seems kind of weird. Maybe SETI decided to make every new user feel at home with an older Joined: date. >;-) Thanks....

L8R....

T'Khasi Time: Monday, 21 March 2005 - 08:07 PM --800 (Pacific Standard Time)

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

> I can only imagine how unmercifully hard all our machines will be hammering
> them for many hours to come, after this extended outage.
> So yeah you're lucky. Enjoy :)

Each project gives BOINC a list of parameters, like how often the clients should connect. There apparently isn't much in the way of a mechanism for the project to adjust those parameters to respond to (anticipated) load.

... and the random back-off isn't quite big enough when there are 200,000 hosts on the project.

They need a bigger back-off when the project is down, and hopefully a way to communicate that to the clients during an outage.

If the back-off was bigger, fewer clients would compete for upload/download/scheduler traffic, and things would ramp back up more gracefully.

---

> > I can only imagine how unmercifully hard all our machines will be
> hammering
> > them for many hours to come, after this extended outage.
> > So yeah you're lucky. Enjoy :)
>
> Each project gives BOINC a list of parameters, like how often the clients
> should connect. There apparently isn't much in the way of a mechanism for the
> project to adjust those parameters to respond to (anticipated) load.
>
> ... and the random back-off isn't quite big enough when there are 200,000
> hosts on the project.
>
> They need a bigger back-off when the project is down, and hopefully a way to
> communicate that to the clients during an outage.
>
> If the back-off was bigger, fewer clients would compete for
> upload/download/scheduler traffic, and things would ramp back up more
> gracefully.
>
The backoff time can be over 3 hours which I believe is big enough.
I doesn't take more than 4 or 5 trys on the upload and it is over 3 hours to connect. The downloads have shorter backoff times.
This gives the downloads a little higher priority.

---

Red Bull Air Racing

Gas price by zip at Seti