What reason was given for the failure?

I wonder Jim, when they get pulled from the server does that stop those already gone out from being re-issued when they come back so the ones that have already gone out have to complete however many failures it is they need before they leave the system?

Roger that, Dave. Unless there is a gimmick of which I'm unaware, the failures will haunt us until they exhaust their chances.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Here the same! All hadam3p_eu fail to download.

I've had 15 fail to download, almost all of them have also failed on other machines.

Got 10 new ones, all downloaded ok.

But now they have all gone. Withdrawn again?

Not withdrawn again but all taken. So many thousand computers with an average of I guess 3 or more processors each all waiting for work...........

In this last batch of Windows work units (issued around 0030 UTC 15 Nov), I had 11 download failures and 22 successes. The 11 seemed to have all failed on 3 machines so they are "done".

---

Finished the AFR WU from Nov 13 without a problem.
I've now picked another AFR WU from Nov 13. This one crashed on another computer with a "Model crashed: REPLANCA: PP HEADERS ON ANCILLARY FILE DO NOT MATCH" when it was almost finished. Let's see if I can finish it :-)


PS: Does the server choose computers with a good track record to send models that already crashed, or is it random? It's not the first time I finish models that crash in other computers. Or am I just "lucky" to receive a resent model with 12000 computers looking for work?

---

Professor Desty Nova
Researching Karma the Hard Way

1)
REPLANCA: (etc) is a problem with mismatched files in the data set.
You'll crash it at the some spot with the same error.

2) You're just lucky. Or, in this case, unlucky. :)

But having a good RECENT track record for completing models is what BOINC looks for when asked to get more work. Otherwise you end up with only one download a day, until your computer shows that it's fit and healthy again. :)

Was there a batch of w@home 2 tasks and if so how big was it? I saw (briefly) 9 tasks in the server status page. They were got minutes later. Was this a batch, or was it just a group of reissues? Such a batch would be gone as fast as the server could deal them (unless it was enormous) considering how many starving computers there are attached to this project.

---

My guess is re-issues.
Also my guess is late Friday, UK time. It ALWAYS seems to be late Friday. :)

It's not just no Windows tasks now. There haven't been any tasks at all available for several days now.

It would be sad if people gave up on CPDN and crunched something else instead. However what people are seeing at the moment is a dead project. My worry is that people will start to vote with their feet - so when tasks are eventually available, fewer people will be interested any more. Is Les seriously suggesting that there will be nothing until Christmas? I hope not. (News and Announcements 2.)

This year is predicted to be the warmest for thousands of years. An understanding of the climate has never been more important.

---

Brian

A Christmas.

It wasn't this last Friday, so perhaps next?

Maybe we will get something soon... On the front page, in "Recent CPDN Submissions", there are several tests in the last days...

---

Professor Desty Nova
Researching Karma the Hard Way

If they are preparing new work why don�t they fix the beta site so we can test them? It is still D.O.A.

---

1) Testing finished.
2) Present beta has been having a severe tantrum, and is refusing to be fixed.
3) Probably no more beta, at least in it's present form. More info when available.

One beta left crunching on my linux box, 2% through uploading a bunch of zip files and trickle ups to Andy. Over a gig is not going to be quick with my connection!

Hmm, if the beta failed, doesn't sound promising for "Christmas", whenever it was supposed to happen. Unless of course I misunderstand the implications...

Some beta tasks have gone out on the main site however. The problem is that if they go to those who, "set and forget" and then fall over, the information needed to fix them doesn't get back to the testers. There are plans for an laternative testing structure I understand which will involve pruning out the set and forget people of whom there seemed to be a few on the beta site.

And just because there is now no "public" beta site, doesn't mean that people in the OeRC can't do in-house testing.

Everybody Relax.

Remain calm.

Go Ommmmmmmmmm.