I've had a few Error in download's recently, reporting as such in a couple of minutes, but this one has got stuck in transit trying to download 6 files since yesterday. The wingman started it in Nov but it took until 4 Feb to report its download error so I suspect that's what's going to happen here. I would have thought that there would have been a shorter deadline to complete the download rather than the full task deadline. The status column already declares No Resubmission and I've just noticed its deadline as Aug 2022 so obviously there's something not right. All very similar to JIM's experience below.
Event log shows:
05/02/2013 19:38:51 | climateprediction.net | Temporarily failed download of HadISST_SST_N96_1990_12_1993_01f.gz: connect() failed
05/02/2013 19:38:51 | climateprediction.net | Backing off 20 min 4 sec on download of HadISST_SST_N96_1990_12_1993_01f.gz
05/02/2013 19:38:54 | | Project communication failed: attempting access to reference site
05/02/2013 19:38:56 | | Internet access OK - project servers may be temporarily down.
but there are no other reports on the boards of servers being down so I suspect the files are on a server which has now been replaced but the files have not been moved over to that replacement.

It gets one more manual retry then it's for the chop.

I aborted each of the 6 stuck transfers so the task declared itself as download failed and then phoned home to report as such. I think a simple abort may have left some unwanted bits and pieces lying about so this way it tidied up properly on its way out.

[several edits for bad grammar, spelling etc.]

Hi Ray

That model that you linked was from June 2011, so it may be the BOINC bug again.
After a very long time, any wu that hasn't met it's max # of error/total/success tasks target gets re-sent by the server. There's a fix, but it requires a server upgrade, which is part of what's going on at the moment.

The worst lot for doing this seems to be the SAF variety.

---

Backups: Here

I've trawled through about 20 of these PNW workunits. There seems to be a whole batch of PNW models that failed to download properly when first sent out and are now being reissued with a 2022 deadline. I also have one with six files stuck in download transfer. Quite a few members with reliable computers who received models from this batch back in November aborted them.

I've asked the programmers about mine. In any case, as far as I can see the files stuck in transfer never even start to download ie they're not using up anyone's bandwidth allowance.

BTW, while trawling through workunits I came across a computer with no owner listed. I don't think I've ever seen this before. If members don't want their name to appear they're listed as Anonymous.

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1266585

I'm sure this is completely irrelevant to the matter of unsuccessful downloads.

---

Cpdn news

More download failures. Hadam3p_pnw_6w17_2008_1_007819365_2 failed to download successfully.

One interesting thing, there don�t appear to be any files stuck in the transfer tab. At least this batch of bad WU�s cleanup after themselves.

P.S. It would be very nice if the Scientists would drop in some more of the Hadam3p_ANZ WU's. Most people haven't had a chance to get a good look at them yet.

---

Jim, yesterday I had a PNW model that failed to download properly in exactly the way you describe.

I wonder whether we should advise members in the News thread to check whether they have the ANZ model type selected in their project preferences. In my account I found that it was deselected by default.

---

Cpdn news

Thanks Mo, I didn't have the NZ models enabled so I guess the majority of CPDN users who don't browse the fora regularly don't either!

Mo's idea of using the News thread to advise the crunchers to enable hadam3p_anz Wu is a good one. When I checked my �preferences� the day they moved over from the beta site, I noticed that ANZ is not selected by default.
[/quote]

A News post will only be a good idea when the current testing with small batches is completed.

---

Backups: Here

Sorry, I didn�t know it was only in limited release.

Early this morning (Eastern time zone, USA) I downloaded an hancm3n, and what do you know, it downloaded completely and is presently running.

Maybe someone dumped a big batch onto the server, I don�t know. By the time I checked �server status� it showed zero available, but, they would go fast with thousands of hungry computers out there looking for work. Nice to know that there are some good ones in the hopper not just the defective retreads.

---

This is the model:

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=15617958
You downloaded it within 4 seconds of its creation.

---

Cpdn news

FOUR SECONDS!!! That has to be a record.

Having waited ages for a work unit, I now have a permanent HTTP error:

24/03/13 23:29:11 | climateprediction.net | Sending scheduler request: To send trickle-up message.
24/03/13 23:29:11 | climateprediction.net | Requesting new tasks for CPU
24/03/13 23:29:15 | climateprediction.net | Scheduler request completed: got 1 new tasks
24/03/13 23:29:18 | climateprediction.net | Started download of hadam3p_eu_9rbj_1977_1_007867263.zip
24/03/13 23:29:18 | climateprediction.net | Started download of atmos_9rbj_1977_1_007867263_0.gz
24/03/13 23:29:20 | climateprediction.net | Giving up on download of hadam3p_eu_9rbj_1977_1_007867263.zip: permanent HTTP error
24/03/13 23:29:20 | climateprediction.net | Giving up on download of atmos_9rbj_1977_1_007867263_0.gz: permanent HTTP error
24/03/13 23:29:20 | climateprediction.net | Started download of eu_9rbj_1977_1_007867263_0.gz
24/03/13 23:29:21 | climateprediction.net | Giving up on download of eu_9rbj_1977_1_007867263_0.gz: permanent HTTP error

Last successful CPDN download was 22/03/13.

Any suggestions please?

That task will not work. Abort the task (and downloads if necessary) and wait for another one to download.

I don't think it's the fault of your computer or connection. I've looked at the workunit your Hadam EU model belongs to and the adjacent workunits. This batch downloaded correctly as far as I can see when it was first released last year. Only a few models from this batch are being resent. I've found one that has crashed on download with the same error code and messages as yours, but most of the resent models haven't reported back yet.

Models of the Hadcm type appear to be downloading correctly at the moment. And according to the Server Status page, all the servers are shown as up and running.

I don't think I've seen enough of this type of crash to identify a pattern yet that would have to be reported to Andy. We'll have to see whether more models suffer the same fate or this was bad luck, a one-off.

This is the sort of report that can pinpoint problems so thank you. I hope you're lucky and get another model soon. If you get repeated download failures please let us know.

---

Cpdn news

Hadcm models are downloading correctly. I got 2 overnight and both are running fine.

---

Well, I went to Abort the Task/Transfer only to find that it had cleared all by itself, so thank you all for your input but no action necessary on my part. Needless to say, there are no tasks available but that is quite a different problem!

Thanks again.

What ever happened to the downloads, that work unit self-aborted.
If you look at the top of the workunit page here, you'll see that it was created 11 April 2012, so it was re-sent by the BOINC bug, and as such wasn't/isn't needed.

---

Backups: Here

I think that I got some form of phantom download overnight. I believe it was an Hadam3p_pnw. It seems to have failed with a permanent HTTP error. Is this an isolated failure or do we have another batch of bad WU�s?

4/17/2013 1:38:44 AM | climateprediction.net | Scheduler request completed: got 1 new tasks
4/17/2013 1:38:47 AM | climateprediction.net | Started download of hadam3p_pnw_brgv_1998_1_007921902.zip
4/17/2013 1:38:47 AM | climateprediction.net | Started download of ic19610106_10_N96.gz
4/17/2013 1:38:56 AM | climateprediction.net | Giving up on download of hadam3p_pnw_brgv_1998_1_007921902.zip: permanent HTTP error
4/17/2013 1:38:56 AM | climateprediction.net | Started download of xaclfa.start.0000.gz
4/17/2013 1:39:01 AM | climateprediction.net | Finished download of ic19610106_10_N96.gz
4/17/2013 1:39:01 AM | climateprediction.net | Started download of so2dms_N96_1998_12_2001_02.gz
4/17/2013 1:39:12 AM | climateprediction.net | Finished download of so2dms_N96_1998_12_2001_02.gz
4/17/2013 1:39:12 AM | climateprediction.net | Started download of dchaba.start.pnw.b.0000.gz
4/17/2013 1:39:24 AM | climateprediction.net | Finished download of dchaba.start.pnw.b.0000.gz
4/17/2013 1:39:24 AM | climateprediction.net | Started download of HadISST_SI_N96_1998_12_2001_01f.gz
4/17/2013 1:39:30 AM | climateprediction.net | Finished download of HadISST_SI_N96_1998_12_2001_01f.gz
4/17/2013 1:39:30 AM | climateprediction.net | Started download of HadISST_SST_N96_1998_12_2001_01f.gz
4/17/2013 1:39:53 AM | climateprediction.net | Finished download of xaclfa.start.0000.gz
4/17/2013 1:40:00 AM | climateprediction.net | Finished download of HadISST_SST_N96_1998_12_2001_01f.gz
4/17/2013 2:39:21 AM | climateprediction.net | Sending scheduler request: To fetch work.
4/17/2013 2:39:21 AM | climateprediction.net | Reporting 1 completed tasks

---

It looks like a spurious reissue of a year-old Task. No harm done because of the crashes, except to your download bandwidth. You were stuck with three of them today, Jim. Regrets.

(Database access is a bit flaky [technical term!] at the moment. Work Unit access is okay but not Task access.)

---

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