Need about a million tasks to keep us busy for any length of time.

There are 8.9 million Tasks queued up right now waiting to be processed. It had been 10 million+ for quite a while.
However until they fix their server systems so that the amount Ready to send is more than 0 for more than an hour or two each day (it used to be 5,000) and the Assimilators can keep up, there will continue to be millions of Tasks waiting to be processed for some time to come.


Then of course fixing up their DNS & download sever issues would then allow even more people to process work without having to screw around with what should be completely un-necessary Hosts file settings.
But considering the download server has been an issue for most of this year (if not longer), i don't hold out any real hope of anything being fixed.


Give it another month or so and there will be so many server issues that Researchers will probably keep sending out work to be processed, but have no idea why they aren't getting any results beck- because no one can download any, and if they could they still can't get any allocated to be downloaded.



The stupidity of this situation is reaching mind boggling levels of ridiculousness.

---

Grant
Darwin NT

I had added one of the lines (I did not see two lines at the time) to Windows 10 hosts file and it had solved a problem.
However, I was getting the 'feeder not running' error and noticed My host file was empty again. Added both lines this time.

Not sure if it is some security software that I need to let it know about the entries or a Windows update that reset the hosts file.

Also, this is not an issue with my other computers. no hosts update needed on 2 with Windows 10 and an old Win 7 that is still running.
(Finally stopped using Win95 several years ago but still sort of miss XP.)

It looks like everything of yours has uploaded now.
You're not the first to report your hosts file going blank. It's the first place to check if you're getting that feeder error again.
And like you, 3 of my devices have never had a problem, but one won't behave itself without these edits.
I don't think anyone knows why some are susceptible and some aren't tbh.

Are we running the same version of windows? I have 64bit pro and my version is 22h2

Need about a million tasks to keep us busy for any length of time.

There are 8.9 million Tasks queued up right now waiting to be processed. It had been 10 million+ for quite a while.
However until they fix their server systems so that the amount Ready to send is more than 0 for more than an hour or two each day (it used to be 5,000) and the Assimilators can keep up, there will continue to be millions of Tasks waiting to be processed for some time to come.


Then of course fixing up their DNS & download sever issues would then allow even more people to process work without having to screw around with what should be completely un-necessary Hosts file settings.
But considering the download server has been an issue for most of this year (if not longer), i don't hold out any real hope of anything being fixed.


Give it another month or so and there will be so many server issues that Researchers will probably keep sending out work to be processed, but have no idea why they aren't getting any results beck- because no one can download any, and if they could they still can't get any allocated to be downloaded.



The stupidity of this situation is reaching mind boggling levels of ridiculousness.

In Robetta, there are things called structure prediction and domain prediction. Whats the difference?
And then some of the tasks are listed with username Rosettafold, who are the rest?

I had added one of the lines (I did not see two lines at the time) to Windows 10 hosts file and it had solved a problem.
However, I was getting the 'feeder not running' error and noticed My host file was empty again. Added both line this time.

Not sure if it is some security software that I need to let it know about the entries or a Windows update that reset the hosts file.

Nothing should be playing with your Hosts file, however some sort of AV/Antimalware programme may be configured to "protect" it , and once it notices that it's changed, it reverts it back to whatever it considers to be the correct Hosts file.
No Widows update that i can think of has ever had anything to do with the Hosts file.

---

Grant
Darwin NT

In Robetta, there are things called structure prediction and domain prediction. Whats the difference?
And then some of the tasks are listed with username Rosettafold, who are the rest?

Not a clue.

---

Grant
Darwin NT

In Robetta, there are things called structure prediction and domain prediction. Whats the difference?
And then some of the tasks are listed with username Rosettafold, who are the rest?

Not a clue.

Well that is why I question the "millions" of tasks waiting. Because I think some of them are processed with different machine(s) than ours. Otherwise we would have work coming out our ears.
I could never really understand Robetta. Is it a server or is a processor or a work generator based on proteins they think need to be examined.

And with 78,000 tasks given to 14,340 systems containing an average of how many cores each....the work they put out is gone in a heartbeat. Maybe they run out the queue for a specific protein and then up load the rest.

Nothing should be playing with your Hosts file, however some sort of AV/Antimalware programme may be configured to "protect" it , and once it notices that it's changed, it reverts it back to whatever it considers to be the correct Hosts file.
No Widows update that i can think of has ever had anything to do with the Hosts file.

On my Windows 8.1 system I had a similar problem of host file being blank or reverting to the Windows default after applying some custom configs...

Discovered it was Microsoft Windows Defender Antivirus protection thinking it was a 'virus' making changes. Solve problem by EXCLUDING file from being auto scanned.

.dupe.

.dupe.

Well that is why I question the "millions" of tasks waiting. Because I think some of them are processed with different machine(s) than ours. Otherwise we would have work coming out our ears.

As i have pointed out multiple times- we have millions of Tasks Queued up to be processed- as shown on the Home page.

But there are issues with the Servers, so they are not being moved to the Ready to send so we can't be allocated them so we can't process them, which is why there are still so many of them sitting there after over a week of increased system issues.
If they were being processed by other systems, they would all be gone by now. If we were able to process them, they would all be gone by now.

Along with the Download server issue, and the IPv6 issue, there are also Assimilator issues, along with the problem with the Feeder not feeding the Ready to send buffer.

---

Grant
Darwin NT

Well that is why I question the "millions" of tasks waiting. Because I think some of them are processed with different machine(s) than ours. Otherwise we would have work coming out our ears.

As i have pointed out multiple times- we have millions of Tasks Queued up to be processed- as shown on the Home page.

But there are issues with the Servers, so they are not being moved to the Ready to send so we can't be allocated them so we can't process them, which is why there are still so many of them sitting there after over a week of increased system issues.
If they were being processed by other systems, they would all be gone by now. If we were able to process them, they would all be gone by now.

Along with the Download server issue, and the IPv6 issue, there are also Assimilator issues, along with the problem with the Feeder not feeding the Ready to send buffer.

Hah! And they call this a cutting edge project.
It took GPUgrid to move and sort out its problems in 3 weeks to a month and they are younger than this project.
This is beginning to sound like a poor WCG after they moved.
The real question is does Baker Lab care anymore or do they not have enough staff or what?
That they ignore emails says a lot.
How long have we had this problem of sporadic tasks and other time frames of no work? Years I think.
So where the he11 is everyone?

Lines added to host, still receiving message: "feeder not running".

This is beginning to sound like a poor WCG after they moved.

The real question is does Baker Lab care anymore or do they not have enough staff or what?

That they ignore emails says a lot.
How long have we had this problem of sporadic tasks and other time frames of no work? Years I think.

So where the he11 is everyone?

It's been several years of seeing project "owner" David Baker post anything and as you can see by browsing the forums, no one else is tending to the ship as it sinks deeper into the abyss.

Thank you for that info. I disabled IPv6 and got two Rosetta tasks instantly. (^:

---

Well that is why I question the "millions" of tasks waiting. Because I think some of them are processed with different machine(s) than ours. Otherwise we would have work coming out our ears.

As I have pointed out multiple times- we have millions of Tasks Queued up to be processed- as shown on the Home page.

But there are issues with the Servers, so they are not being moved to the Ready to send so we can't be allocated them so we can't process them, which is why there are still so many of them sitting there after over a week of increased system issues.
If they were being processed by other systems, they would all be gone by now. If we were able to process them, they would all be gone by now.

Along with the Download server issue, and the IPv6 issue, there are also Assimilator issues, along with the problem with the Feeder not feeding the Ready to send buffer.

Hah! And they call this a cutting edge project.
It took GPUgrid to move and sort out its problems in 3 weeks to a month and they are younger than this project.
This is beginning to sound like a poor WCG after they moved.
The real question is does Baker Lab care anymore or do they not have enough staff or what?
That they ignore emails says a lot.
How long have we had this problem of sporadic tasks and other time frames of no work? Years I think.
So where the he11 is everyone?

Greg, you're working yourself up into one of your tantrums again, which is how you ended up departing for a few years until recently.
I know you have high expectations from the projects you give your time to, which is why I tried to make clear to you when you returned in January that while task availability was still very far from perfect, being here was workable with a little patience and effort. Since then we've had the security cert expiry issue, the IPv6 issue and the go-slow of bwsrv1 - all of which, I completely accept, are highly annoying, but with the ingenuity of other users here we've found our way through it.
That's certainly not ideal and certainly doesn't meet your high expectations, but most of us accept this is what this project has become without constantly harking back to a past that's long gone.

This project certainly is cutting-edge, but for the science it works on, not ever for its forum-responsiveness or for its server management performance.
Rosetta has always had a very small team dealing with its admin - that's your 'everyone'.
Boboviz posts regularly on the new features they're constantly developing. That's what they're focussing on. It would be ridiculous if they distracted themselves from their nobel-prize-winning research to monitor their forums or servers.
In the short time I did have someone's attention, it was made very clear to me the tail does not wag the dog at this project, nor ever will it.
That's because what they care about is 99.99% their research and 0.01% their overhead, which is what we are.
That's not the same as what happened at WCG, where IBM effectively abandoned it to a small group who hoped for much more than they were ever capable of delivering.

If you don't want to continue to offer your time here, that's up to you. Your expectations won't realistically be met except through the regular expertise and generosity of other users here.
But istm the worst of all worlds is understanding/accepting that while simultaneously whining about it several times a week. That gets us nowhere.

---

Well that is why I question the "millions" of tasks waiting. Because I think some of them are processed with different machine(s) than ours. Otherwise we would have work coming out our ears.

As I have pointed out multiple times- we have millions of Tasks Queued up to be processed- as shown on the Home page.

But there are issues with the Servers, so they are not being moved to the Ready to send so we can't be allocated them so we can't process them, which is why there are still so many of them sitting there after over a week of increased system issues.
If they were being processed by other systems, they would all be gone by now. If we were able to process them, they would all be gone by now.

Along with the Download server issue, and the IPv6 issue, there are also Assimilator issues, along with the problem with the Feeder not feeding the Ready to send buffer.

Hah! And they call this a cutting edge project.
It took GPUgrid to move and sort out its problems in 3 weeks to a month and they are younger than this project.
This is beginning to sound like a poor WCG after they moved.
The real question is does Baker Lab care anymore or do they not have enough staff or what?
That they ignore emails says a lot.
How long have we had this problem of sporadic tasks and other time frames of no work? Years I think.
So where the he11 is everyone?

Greg, you're working yourself up into one of your tantrums again, which is how you ended up departing for a few years until recently.
I know you have high expectations from the projects you give your time to, which is why I tried to make clear to you when you returned in January that while task availability was still very far from perfect, being here was workable with a little patience and effort. Since then we've had the security cert expiry issue, the IPv6 issue and the go-slow of bwsrv1 - all of which, I completely accept, are highly annoying, but with the ingenuity of other users here we've found our way through it.
That's certainly not ideal and certainly doesn't meet your high expectations, but most of us accept this is what this project has become without constantly harking back to a past that's long gone.

This project certainly is cutting-edge, but for the science it works on, not ever for its forum-responsiveness or for its server management performance.
Rosetta has always had a very small team dealing with its admin - that's your 'everyone'.
Boboviz posts regularly on the new features they're constantly developing. That's what they're focussing on. It would be ridiculous if they distracted themselves from their nobel-prize-winning research to monitor their forums or servers.
In the short time I did have someone's attention, it was made very clear to me the tail does not wag the dog at this project, nor ever will it.
That's because what they care about is 99.99% their research and 0.01% their overhead, which is what we are.
That's not the same as what happened at WCG, where IBM effectively abandoned it to a small group who hoped for much more than they were ever capable of delivering.

If you don't want to continue to offer your time here, that's up to you. Your expectations won't realistically be met except through the regular expertise and generosity of other users here.
But istm the worst of all worlds is understanding/accepting that while simultaneously whining about it several times a week. That gets us nowhere.

I figured out their pattern now. It seems they run a protein and get all the results and then put the next one out. At least that is how it looks now.
As for proteins in Robetta, I randomly sampled a few looking at 'user input' and I found this:

Modeling method CM only
Superimpose partial threads before hybrid sampling.
Sample 1 model.

Which differs from
Modeling method RoseTTAFold only which I believe are all things we process.

As for my 'rant', how do you expect your science to get done if you don't keep your machines up and running properly? I run Denis@home and his project is significantly smaller. But he keeps an eye on his machines and has his team quickly fix any problems. So that is all I am pushing. Keep the machinery running. Forum posts..well if you have a fubar or semi fubar problem, then tell us what it is and how your fixing it. Doesn't have to be a paragraph, just a line or two. Communication helps a lot. Its just bizarre are well established project like this has so many issues. But I guess thats the way it will be now. So whatever...
If they go offline for a long time, then I go away and check back later like now.
But for now its slow and steady so I just let it run in the background.

Boboviz posts regularly on the new features they're constantly developing. That's what they're focussing on. It would be ridiculous if they distracted themselves from their nobel-prize-winning research to monitor their forums or servers.

No, what is ridiculous is having a public project, with all sorts of server-side issues, with no support at all from those responsible for the project.
It does not reflect well on Baker Labs at all.

It doesn't need 24/7/365 monitoring; just to check on it every few days and take care of any issues that arise.

That's because what they care about is 99.99% their research and 0.01% their overhead, which is what we are.

Cosidering that we are responsible for providing them and other researchers with results for their research, that is a ridiculously short sighted point of view, and if it were true then they should just close down Rosetta and be done with it.
If they want results from the work they send out, then they need to make sure that the infrastructure is functional enough to make that possible.

If doing so takes up too much time and effort on their part- better to kill it off once and for all than just have it limp along, driving away many of those that were regular contributors, and driving away any new people that might happen to come along and be completely put off by the need to do all sorts of jumping through hoops just to be able to get & return work.
It's one thing to tweak overall BOINC compute settings, and project specific settings when attaching to a new project, but having to stuff around with Hosts file entries just to be able to do work & return it really is ridiculous.

They should fix it up, or just wind it up all together (officially- not like what they've done with Ralph).
(Having said that, for almost 2 hours there have been more than 1000 Tasks Ready to send, which has been unheard of for the last 2 weeks. There has also been a huge spike in Work in Progress, and there is no Assimilator backlog . *fingers crossed*).

---

Grant
Darwin NT

Grant, you said what I was feeling but did not know how to express. Double thumbs up to you!

(Having said that, for almost 2 hours there have been more than 1000 Tasks Ready to send, which has been unheard of for the last 2 weeks. There has also been a huge spike in Work in Progress, and there is no Assimilator backlog . *fingers crossed*).

Ignore that- it's still broken.

---

Grant
Darwin NT

(Having said that, for almost 2 hours there have been more than 1000 Tasks Ready to send, which has been unheard of for the last 2 weeks. There has also been a huge spike in Work in Progress, and there is no Assimilator backlog . *fingers crossed*).

Ignore that- it's still broken.

bahahaha...they will get to it when it reaches 100,000 gees...