I keep adding and removing Einstein@home to my BOINC client (I remove it because it hogs my CPU/GPU).
I was advised to reduce my cache size to fix the problem of Einstein@Home always running in high priority mode ...
So I have set 0.4 days as my cache size and 0.1 days additional work buffer (in network settings).
This should make sure I have no more than half a days worth of work units WUs in my BOINC client tasks list - right ?
WRONG
As soon as I added the project I got 12 WU/tasks each with an estimated 8 hour run time (4 days worth of work).
So I went and set "No New Tasks" for this project (then I hit the update button to make sure the Einstein@Home server gets the message).
Well the Einstein@Home got the message - and I got an ADDITIONAL 12 WUs/tasks !!!
So I thought xxxx you I'm aborting the last 12 WUs/tasks you sent me after I specifically told you NO NEW TASKS
I went to look at my task list in BOINC (to make sure I have just the original 12 WUs/tasks - which was too many as it was considering my cache size, but I though what the hell I can suspend them later if I want).
What do I see immediately after I aborted the 12 additional WUs/tasks ? I see MORE Einstein@Home tasks - status DOWNLOADING EVEN MORE !!!. Ended up with 24 AGAIN.
So I decided to post this in the forum (but before posting I decided to once again count the number of Einstein@Home WUs/tasks to make sure my numbers are correct) ...
I have a total of 60 (SIXTY) !!! Einstein@Home WUs/tasks in my BOINC client now - so in the time it took for me to almost finish writing this post I received another 36 Einstein@Home WUs/tasks (remember my prefernces in BOINC and online have a cache size of just half a day).
I am treating Einstein@Home as malware from now on, because project administrators do not respect BOINC preferences.
If BOINC was an email client, Einstein@Home would be treated as SPAM and blocked by a SPAM filter.
Because of this I am making a formal complaint to the BOINC administrators (to have Einstein@Home removed from BOINC's project list, and if possible to prevent Einstein@Home from adding to the BOINC client in other ways).
I have screen recordings and witnesses using other machines that have recorded the same behaviour.
I will find out who is responsible for enforcing compliance with BOINC Terms of Use by project administrators, and will report it to the responsible authorities to make sure Einstein@Home project administrators stop this blatant misuse of the BOINC software, it's licence terms and conditions (I am prepared to take legal action against these misleading and deceptive conduct if the above remedies don't work).
Enough is enough.
Copyright © 2024 Einstein@Home. All rights reserved.
No New Tasks Set in BOINC - STILL Getting New Tasks
)
Yay - got another 10 Einstein@Home WUs/tasks since my last post (less than an hour ago) ... this is a xxxx joke !!!!
BTW my settings in BOINC are STILL "no new tasks" for Einstein@Home - but what do they care what my settings/preferences are huh ???
I had a different, but
)
I had a different, but similar problem some time ago.
I finally had to resolve it by removing the project and re-adding it.
It's nothing anyone is doing on purpose.
RE: I had a different, but
)
Hi tbret - I did just that (if you read the FIRST line of my first post).
This isn't an "intermittent problem", it is a systematic abuse of the BOINC program (and all users who attach an Einstein@Home project).
If you don't believe me - do what I did (remove the project then add it) and see what happens.
They are breaking the law (I just have to get my solicitors/lawyers to determine the jurisdiction in which I can prosecute them for it - and what is involved).
The fact that their actions are unlawful are clear (according to my legal advisors) it is just a case of proving it (in the right jurisdiction).
RE: I will find out who is
)
Humm... do you realize that the components which decide on work request are BOINC components not Einstein components?
RE: I keep adding and
)
This project has old Boinc scheduler code, where the scheduler will resend existing tasks allocated to your computer even with NNT set,
more resent scheduler code will only resend lost work if the client asks for work,
Your host has at present 159 tasks allocated to it already:
In progress tasks for computer 11635427
Boinc is only resending those existing tasks, it is not sending new tasks:
http://einstein5.aei.uni-hannover.de/EinsteinAtHome/host_sched_logs/11635/11635427
Claggy
This same thing has happened
)
This same thing has happened to me when turning off new tasks... I still get more tasks.
I have a question: I have 29
)
I have a question:
I have 29 "in progress" tasks at the moment. If I were to set Einstein at Home to "no new tasks", it would still finish all of the "in progress" tasks before it stopped doing work.
Is that correct?
Joshua_48 wrote:This same
)
A couple of points.
It's not helpful to resurrect a four year old 'dead' thread where you haven't provided any information to support your guess that this is the 'same thing'. There are a lot of incorrect assertions and misinformation about what is going on (by the OP) and where the 'blame' should be attributed. Needless to say, there will be a logical explanation and a simple solution to whatever your problem happens to be. You just need to describe what you have done and what has happened. Since your computers are hidden, people wishing to help will be unable to get the sort of information that would help pin down the problem.
You would be far better off ignoring that rambling diatribe that you have 'discovered' and start a new thread on the 'Problems' board, explaining exactly what your problem really is. When you describe the circumstances around your problem, it would be helpful if you would mention the host ID that has been assigned to your machine. That would allow a helper to look at scheduler logs that document the transactions between your host and the scheduler which sends any work requested by your BOINC client.
Until you explain your problem and get a response, just leave NNT (no new tasks) set in BOINC manager and let your BOINC client manage the situation as it is designed to do. Make sure your work cache settings are at nice low values - no more than say 0.5 days in total - preferably in the first setting and zero in the 'extra days'. If you are running out of work, reverse the NNT setting once you have set the 0.5 work cache setting.
EDIT: Your second message arrived whilst I was composing the above. The answer to the extra question is YES.
Cheers,
Gary.
Thanks for your reply Gary.
)
Thanks for your reply Gary. I'm new so I'm still trying to figure out how everything works. I'll start a new post if I have any more questions. Thanks again!
You're welcome! I understand
)
You're welcome!
I understand very clearly that BOINC can be a bit daunting and rather puzzling in some of its intricacies, when people first join. That (in particular) is why I'm trying to ensure that you don't latch on to some of the misconceptions about how things really work that were being complained about in the opening post of this thread.
Here are some simple facts to chew on that weren't understood by the thread starter all those years ago.
This is just a very small subset of some of the 'rules of engagement' if you want to make participating simple and enjoyable. It is by no means complete or comprehensive. One of the nice things about Einstein is that it tends to be quite reliable so there really is little need to have a large store of work on hand. Once you get used to things, you could consider a setting of a day or two, just to protect against a problem over a weekend. There are very few problems that take more than a day or so to be resolved.
I highlighted the word 'new' above for a specific reason. When work is first allocated to your computer it will be brand new work. If you 'lose' work for any reason outside of directly aborting it and reporting it to the project as 'aborted', the project will 'remember' the work that has been allocated to you. For example, lets say something gets screwed up on your host and you reinstall. The next time you contact the project (even with NNT set) the project is quite likely to recognise your machine and will therefore 'know' what you are supposed to have. The project will resend those tasks to you, describing the process as "resending lost tasks".
They are not new tasks, they are 'resends', and this is why NNT doesn't prevent those tasks from being resent to you. If you really don't want them, they can be aborted and reported back to the project as 'aborted' and that is the proper way to ensure they will be gone permanently. Resends are only sent in batches of 12 at a time. If there are more than that, more will arrive with each successive project contact by your client. Einstein uses customised server software based on older versions of the BOINC server code. For that reason, there may well be differences with how other projects handle the resending of lost tasks.
Cheers,
Gary.