Where'd you get that from? I've got 8 Windows and 1 Linux PC and I never get work.

I have some (old) Macs so I only get the Hadley CM3 jobs. These three were sitting in my Transfers tab, trying to be downloaded, for a month. The most previous work before that was from February. There's not a lot to go around from this project. I counted and my two Macs have been assigned 28 tasks since May 18th of last year.

I guess it's a good thing, the scientists (if the server was behaving!) get work done immediately. I'd much rather sign up to 20 projects and see bursts of work from each, instead of the scientists waiting years for a massive batch to get done. I guess we just need more power - I've just upgraded four machines to 96GB RAM to cater for LHC, but I'm still stuck on this pitiful 7Mbit uplink for 2 years because of my telecoms provider being useless. The whole town but three streets upgraded to 1000/220, and we're stuck on 32/7.

That is not the same issue. This issue in this thread is with the download server having a bad SSL cert. What you are showing is that the project has no tasks available.

That is one thing I thought I was showing: no complaints about bad SSL certificates.
The fact that there are no tasks available is true, but not really surprising.

---

That is not the same issue. This issue in this thread is with the download server having a bad SSL cert. What you are showing is that the project has no tasks available.

That is one thing I thought I was showing: no complaints about bad SSL certificates.
The fact that there are no tasks available is true, but not really surprising.

Ouch, I misread that completely. I apologize for any and all snark in my response.

Ouch, I misread that completely. I apologize for any and all snark in my response.

No problem. Apology accepted.

Just a side-effect of non face-to-face live communication. It happens a lot.

It will probably happen again, knowing human nature -- including my own.

---

Face to face doesn't make it any better. The English language is rubbish.

Face to face doesn't make it any better. The English language is rubbish.

The very reason face-to-face is superior to English..I think Latin is better than English. But face-to-face is better because in addition to the words and associated grammar, the affect communicates much more than the words do, and the immediacy permits errors to be detected sooner and corrections to be made much faster.

---

Face to face doesn't make it any better. The English language is rubbish.

The very reason face-to-face is superior to English..I think Latin is better than English. But face-to-face is better because in addition to the words and associated grammar, the affect communicates much more than the words do, and the immediacy permits errors to be detected sooner and corrections to be made much faster.

Agreed, but using English as the base of communication even face to face, people have misunderstandings. What we need is a precise language, but without all the unnecessary nonsense, like distinguishing between less and fewer.

If the program executable files can't be downloaded from the associated url, that's a bit more serious than the missing (task-specific) atmos.gz files that bullschuck reported earlier.

Program executable names and urls should be persistent in client_state.xml - we should be able to diagnose them further by trying to download them via a browser.

It's possible that the new user already had the program files, retained after a previous job. If that's the case, there would be no need to attempt a fresh download.

It's possible that the new user already had the program files, retained after a previous job. If that's the case, there would be no need to attempt a fresh download.

Looks like that is the problem. I too have a failed download with

Sat 20 May 2023 17:58:27 BST | climateprediction.net | Scheduler request completed: got 1 new tasks
Sat 20 May 2023 17:58:27 BST | climateprediction.net | Project requested delay of 3636 seconds
Sat 20 May 2023 17:58:29 BST | climateprediction.net | Started download of hadam4_8.52_i686-pc-linux-gnu
Sat 20 May 2023 17:58:29 BST | climateprediction.net | Started download of hadam4_data_8.52_i686-pc-linux-gnu.zip
Sat 20 May 2023 17:58:31 BST | climateprediction.net | Giving up on download of hadam4_8.52_i686-pc-linux-gnu: permanent HTTP error
Sat 20 May 2023 17:58:31 BST | climateprediction.net | Giving up on download of hadam4_data_8.52_i686-pc-linux-gnu.zip: permanent HTTP error
Sat 20 May 2023 17:58:31 BST | climateprediction.net | Started download of hadam4_se_8.52_i686-pc-linux-gnu.zip
Sat 20 May 2023 17:58:31 BST | climateprediction.net | Started download of hadam4_um_8.52_i686-pc-linux-gnu.zip
Sat 20 May 2023 17:58:32 BST | climateprediction.net | Giving up on download of hadam4_se_8.52_i686-pc-linux-gnu.zip: permanent HTTP error
Sat 20 May 2023 17:58:32 BST | climateprediction.net | Giving up on download of hadam4_um_8.52_i686-pc-linux-gnu.zip: permanent HTTP error

I could probably copy the files across from somewhere else now I know but I wasn't at the computer when it happened. Because I did a clean install of XUbuntu 23.04 I don't have any of the executeables. I have let Andy know.

It also affects IFS executables (those are the only ones I've run on this incarnation of Linux).

client_state.xml gave me the download url as https://www.cpdn.org/applications//oifs_43r3_ps_1.05_x86_64-pc-linux-gnu, and the response was

Not Found

The requested URL /applications/oifs_43r3_ps_1.05_x86_64-pc-linux-gnu was not found on this server.

I've added that to Dave's email to Andy.

The acid test will be when we start getting downloads across the board.

I do not know about across the board, but I notice I got a work-unit (a re-run) very recently. Boinc-manager is predicting a little over 8 days to complete it. It has run 46 minutes so far.

Machine ID: 1511241
Task        22318255
Name 	hadam4h_a05p_200011_5_931_012138767_2

---

But first I need to work out how to copy files across from the guest to the host OS. Copying text works but not files so far. Am looking at various threads on intereweb about this!

Hi Richard,

Thank you very much, this should be okay now.

Best wishes,

Andy

I can confirm the files download via a browser now.
Edit: Some do, there are still a couple of files not downloading for me.

... there are still a couple of files not downloading for me.

In Dave's email to Andy he identified two distinct remaining problems.

1) Files which are still missing from the server - they can't be downloaded at all, either by BOINC or by a browser. There's nothing we can do about these - we just have to wait for Andy to find the files and put them back in the right place.

2) Files which have have been re-instated already, and can be downloaded by a browser - but not by BOINC.

This may be related to a bug I reported some time ago to BOINC. Under Linux, the client leaves behind a zero-length file when a download fails because the source file is missing from the server. This confuses subsequent attempts, because BOINC thinks the file is already present and skips the download - and of course the file fails subsequent integrity checks.

If you have full administrator access to the BOINC client folder on your machine, and you're still getting download errors other than 404 'file bot found', look for zero-length copies of the affected file in the CPDN project folder and delete them.

Reference: https://github.com/BOINC/boinc/issues/4642

This may be related to a bug I reported some time ago to BOINC. Under Linux, the client leaves behind a zero-length file when a download fails because the source file is missing from the server. This confuses subsequent attempts, because BOINC thinks the file is already present and skips the download - and of course the file fails subsequent integrity checks.

The one BOINC was failing to download, I had managed to find a copy but having put it into the climateprediction.net folder in projects, I had neglected to give it permission to run as a program. Would this cause the same problem as a zero byte size file? I suspect it may have been doing so but I will have to wait till another task comes along to check that having changed its permissions resolves the issue.

Admin access, run as a program, I'm glad I use windows!

No, not the 'same' problem, but it might well cause a different problem with the same end result - task failure.

Andy is saying it should be fixed now.

Andy is saying it should be fixed now.

Fihnally got one of the resends and can confirm from my own experience that it is fixed.