If a computer is trashing every thing due to missing 32bit libs, how does one get in contact with the machine's real owner as opposed to Science United which shows up as the owner if they are signed up via that account manager? I have had a quick poke around on the Science United site and not found anything helpful.

At the moment none of those boxes are ones with lots of cores so not a biggie but at some point it is likely to be and given that these computers are more likely to be owned by the set and forget crowd....

Edit: All but one of the linux machines listed on the cpdn site belonging to Science United that had received tasks were trashing them for the above reason.

Andy Bowery is due to speak at the BOINC PROJECTS CALL #6, a week tomorrow. I'd like to hear him raise problems like this (including the inability of Science United users to access project preferences, or to contribute to project message boards) in front of the wider BOINC project community.

Andy Bowery is due to speak at the BOINC PROJECTS CALL #6, a week tomorrow. I'd like to hear him raise problems like this (including the inability of Science United users to access project preferences, or to contribute to project message boards) in front of the wider BOINC project community.

I emailed him about this before posting here. :)

I emailed him about this before posting here. :)

And I saw that before replying here :)

The point is that I think we ought to think about this as a broader, BOINC-wide, problem with Science United. Did you see the user at SU who asked "How do I accept a Friend request at a project?" [that I can't log in to]. No effective reply. David is very proud that SU processes a million tasks a month. SETI does that three times a day.

But people like Andy (nice chap, met him several times at workshops) tend only to think INside the box, in Andy's case CPDN. I'm busy trying to sort out a cross-project problem with Anonymous Platform, which has existed and been talked about in public and private for 9 months. But everybody regarded it as a private, project, problem, not important for my particular project - until it blew up spectacularly at SETI, just before Christmas.

We need to join the dots, better. I can ride shotgun if he'll only broach the subject, but I think he should be the one to introduce it.

I emailed him about this before posting here. :)

And I saw that before replying here :)

The point is that I think we ought to think about this as a broader, BOINC-wide, problem with Science United. Did you see the user at SU who asked "How do I accept a Friend request at a project?" [that I can't log in to]. No effective reply. David is very proud that SU processes a million tasks a month. SETI does that three times a day.

But people like Andy (nice chap, met him several times at workshops) tend only to think INside the box, in Andy's case CPDN. I'm busy trying to sort out a cross-project problem with Anonymous Platform, which has existed and been talked about in public and private for 9 months. But everybody regarded it as a private, project, problem, not important for my particular project - until it blew up spectacularly at SETI, just before Christmas.

We need to join the dots, better. I can ride shotgun if he'll only broach the subject, but I think he should be the one to introduce it.

Makes sense to me. While I have only dabbled in projects other than CPDN which remains my prime focus, I do recognise that it is a much wider problem which I assume applies to those who sign up with other account managers as well as Science United and obviously to other projects than CPDN. And, yes I saw the post on their forums and a few others that have implications for most if not all projects. What I didn't see on their fora was much in the way of answers.

What I didn't see on their fora was much in the way of answers.

I think it's about time that we declared that the Emperor of SU has no clothes.

GASP!
:)

Perhaps, (if it's not there already), some server code that checks each computer against a list for suitability for the project, and blocks it from being attached WITH A MESSAGE AS TO WHY NOT!
So for cpdn, no android, nothing with missing 32 bit libs, etc.

Or some lightweight probe app that tests for 32-bit libs, and sends back a 'blacklist' message if none found - all before committing a substantial model download.

Andy has created #3420, saying "In climateprediction.net we now have account creation disabled through the BOINC Manager". Dies anyone know - either way - whether this has fixed the SU problem, or can SU still create accounts?

[Edit - SU did attach me to CPDN when I ran a test on 25 Nov 2018, so that would come into the problem category. I've got it deselected in preferences.]

What I didn't see on their fora was much in the way of answers.

I always forget I have an account there and am a moderator apparently. But then I tend to ignore those forums as well, as the one time I thought I made an informed answer, David hid my post. Is when I immediately stopped moderating & coming there, as you can see from all the spammers on those forums.

If he doesn't care, why should I?

---

Having started this hare running (Andy says he'll raise it on the Projects call), I thought I'd better run some CPDN - now I have a Linux box to run it on.

First question, what version of Ubuntu do I have? No matter how nicely I ask it, all it will say is "Linux Mint 19.1 Tessa". Didn't like the curses6 incantation for Ubuntu 19.1, but seems to have accepted curses5 for Ubuntu 18.04 / 19.04

Point at it and laugh out loud? That may make it mad. 😂

---

Well, the CPDN task has been running for an hour and a half so far (out of an estimated 12 days). You'd have thought it would have noticed by now if the libs were missing.

"The usual" missing lib, libstdc++.so.6, will show up at about 6 seconds.

Or some lightweight probe app that tests for 32-bit libs, and sends back a 'blacklist' message if none found - all before committing a substantial model download.

So, under disguise "users don't care, for realz" what would be installed is some code that pokes in donor's privacy. Is CVE already issued? Is it GDPR safe?

---

I'm counting for science,
points just make me sick.

The SU issue was raised on the Projects call yesterday by CPDN, by David Wallom (Andy's boss). David Anderson was one of several people not attending or calling in sick form home - there must be a lot of flu around. But the central BOINC team are going to discuss it at their next meeting.

Bernd Machenschalk of Einstein is trying to contact Andy, because he thinks there might be (or ought to be) a way of preventing clients without the 32-bit libs being sent work. But Bernd's email has bounced - does anyone know if Andy has a particularly strong spam filter, or an alternate email address?

I've just emailed Andy and asked him to look at this thread.