BOINC needs some major fixing

Message boards : BOINC Manager : BOINC needs some major fixing
Message board moderation

To post messages, you must log in.

AuthorMessage
Johnny Andrews

Send message
Joined: 30 Nov 05
Posts: 5
United States
Message 1758 - Posted: 10 Dec 2005, 18:32:55 UTC

I am posting this here in the hopes that someone in the BOINC project will see it, since I have not been able to find an email address to send suggestions or comments to. I have been running seti for years, and so I made the switch to BOINC when I found out about it recently.

So far, BOINC has been nothing except one big pain. At first, I could not attach to the seti project, because BOINC could not resolve the url. This was fixed when I threw BOINC away and redownloaded it. I got a newer version of the BOINC Manager (currently 5.2.13 for the Mac). I managed to attach to seti, and got a work unit to work with.

But seti seems to be having serious problems. I finished the first work unit, and after a number of retries, downloaded a second unit to work on. But I still haven't managed to upload the results from the first work unit. And right now, I cannot upload the results from the first work unit or the second work unit, and can't get a thiird work unit to download. I noticed in a message on this board that seti's problems were known, and presumably they will be fixed eventually.

Since seti is temporarily unavailable, I decided to attach to another project to have something else to work on in the meantime. So I attached to einstein and downloaded a work unit from them. But this has demonstrated another management problem which apparently I cannot resolve. The BOINC manager has allocated 50% of its time to seti, and 50% of its time to einstein. And it won't change this allocation until it can connect to the seti server and review the seti preferences.

The capability of modifying time allocations between projects should be stored on the client computer, not dependent on access to a host. There are several reasons for this, the most important illustrated by my current problems.

I do not have a continuous internet connection. I have a dial-up modem. I learned long ago with seti that connections to the Berkeley computer could be very slow on weekdays. So I used seti's "connect now" feature to upload results and download new work units in the late evening or on weekends. I started up my internet connection before telling seti to connect. The BOINC Manager does not appear to have any control over connection attempts. It does not try to coerce a connection through my modem. But it does (from the messages panel) attemp to continuously connect to seti every time it switches to seti as the current project unit, since it has only the two completed work units for seti. And since, obviously, with no modem connection, it cannot resolve to seti, it sits there until the connection attempt times out before switching back to einstein. Which means that nothing is getting done.

The BOINC Manager should not be attempting to connect to a project host unless a modem connection is detected! And, in the current circumstances, if the time allocation preference was stored in my (client) computer, I could tell BOINC to work continuously on the einstein project until seti gets its problems fixed and becomes accessible.
ID: 1758 · Report as offensive
Michael Roycraft
Avatar

Send message
Joined: 24 Nov 05
Posts: 129
United States
Message 1764 - Posted: 10 Dec 2005, 19:40:30 UTC
Last modified: 10 Dec 2005, 19:42:56 UTC

Mr. Andrews,

As you have not asked any questions here, I will not provide you with any answers. If you'd like to find out what is going on with Seti, why don't you take some time to read the news on their website, particularly Technical News.

SetiClassic people have had over 1 year to make the switch to BoincSeti, and were clearly told that SetiClassic's days were limited. Many made the changeover when they saw the writing on the wall, found new opportunities to contribute, and are very satisfied. Others, in a state of denial, dragged their feet to the bitter end. Most of these last have gracefully adapted to the newer technology. And then there are the rest...

These are like someone who is given a Cadillac after only riding a bicycle. He can't understand that it takes a key to make it work, nor how to adjust the seat or what in the world Air Conditioning is all about. Especially he has a problem with those doggone pedals on the floor - he can't seem to make them crank around in circles, so obviously the Cadillac is broken, must be a piece of junk, right?

He's had the Cadillac for a week now, still trying to make a bicycle out of it, and thinks that he is now an expert on Cadillacs, so he's going to take a trip to the dealership and tell them all about how to build a car.

Sounds absurd. If you want to learn something about a Cadillac, ask a dealer.

Regards,

Michael
"The arc of history is long, but it bends toward Justice"
ID: 1764 · Report as offensive
Paul D. Buck

Send message
Joined: 29 Aug 05
Posts: 225
Message 1768 - Posted: 10 Dec 2005, 20:52:40 UTC - in response to Message 1758.  

Since seti is temporarily unavailable, I decided to attach to another project to have something else to work on in the meantime. So I attached to einstein and downloaded a work unit from them. But this has demonstrated another management problem which apparently I cannot resolve. The BOINC manager has allocated 50% of its time to seti, and 50% of its time to einstein. And it won't change this allocation until it can connect to the seti server and review the seti preferences.

In the projects tab select the SETI project and select suspend. If you have work to complete, use no new work, abort the downloads (if any) then when the work is done select suspend.
ID: 1768 · Report as offensive
Bill Michael

Send message
Joined: 30 Aug 05
Posts: 297
Message 1769 - Posted: 10 Dec 2005, 21:09:20 UTC - in response to Message 1764.  
Last modified: 10 Dec 2005, 21:10:49 UTC

As you have not asked any questions here, I will not provide you with any answers.


@Michael, thank you for an EXCELLENT response. As I was reading Mr. Andrews posting, I was mentally going "yes... we know that... no, that's not how it works... that has been answered 9078685 times... you aren't doing that correctly... that's what you told it to do... so change the preference..." It was quite amusing to realize that there was not ONE thing in his posting that if asked, I could not answer off the top of my head, having done so many times before. Every "improvement" wouldn't be, every "problem" isn't. The entire posting is a result of a simple lack of knowledge, easily remedied by studying a bit, or asking questions.

Mr. Andrews is quite obviously a very intelligent individual, and HAS found many things about BOINC that did indeed, at the time they were decided on, cause discussion before being proven to be best the way they are today. Hopefully he will take some time to ask questions, and read prior board postings, and the Wiki, and then this ability and intelligence will be able to locate REAL problems, and make REAL recommendations for improvement. He just needs to "learn how to drive" first.

But your "as you have not" line, that I'm saving as a scrap, ready to be used... truly classic.

ID: 1769 · Report as offensive
Profile Stefan Ledwina
Avatar

Send message
Joined: 25 Nov 05
Posts: 55
Austria
Message 1770 - Posted: 10 Dec 2005, 21:52:47 UTC

Michael this was a wonderful response! :-)

Greetings Stefan
ID: 1770 · Report as offensive
Profile Tigers Dave

Send message
Joined: 24 Dec 05
Posts: 52
United States
Message 2205 - Posted: 24 Dec 2005, 5:09:37 UTC

Allow me to preface my comments with four disclaimers.

1. First of all, like Johnny Andrews, I am a BOINC newbie. I tried it a while ago on one of machines and didn't like it. Over the last two weeks, as SETI Classic went through the final stages of shutdown, I began installing BOINC on my farm of Macs.

2. Also like Johnny, I have run into some problems installing and setting up BOINC. I have several Macs that do not natively support OS X. Thus, moving them to BOINC has forced me to obtain, install, and configure a third-party patch to enable OS X installation. This process has been difficult and not entirely successful; two of my six PCI PowerMacs still can not run OS X (and hence, BOINC). There are no options whatsoever for installing OS X on five of my other PowerMacs. A more troubling problem was that two of my eMacs that do natively support OS X suffered irrevocable disk failures within 24 hours of installing BOINC. I continue to hope that these were simply unlucky coincidences.

3. As a result of my struggles, I have lurked and solicited advice on BOINC and XPostFacto message boards. Thankfully, participants have been kind, generous, and encouraging.

4. I am uncomfortable with the tone of Johnny Andrews' post. It seems that he wants to complain more than solicit advice.

Nonetheless, I am a bit dismayed with the tone of some (but not all) of the responses to Johnny Andrews' post. Obviously the posters of these responses have tremendous experience and insights into the workings of BOINC. Thus, it is disappointing that instead of trying to help Johnny, they have opted to demean his efforts to make the transition from SETI Classic to BOINC. This type of attitude is what kept many of us who were running SETI Classic from transitioning to BOINC until we absolutely had to and this attitude will undoubted keep some people from making the transition at all.

Please keep in mind that the goal of these projects is to further the science as rapidly as possible. In order to do that, we need as many CPUs as we can muster. Don't discourage those from participating - it goes against the spirit of the project.

Happy holidays and happy crunching!

David Riese

ID: 2205 · Report as offensive
Paul D. Buck

Send message
Joined: 29 Aug 05
Posts: 225
Message 2207 - Posted: 24 Dec 2005, 11:27:45 UTC

David,

Good disclaimers. Some comments on them.

It is sad when we move on in the software world and older computers are left by the wayside as the platform is not supported by the latest systems. But, it was a resources decision and sad to say there is little we can do to alter the logic. I do hope you can get your systems up with out much more difficulty.

We do try to be kind, generous and encouraging. And #4 *IS* the crux of the matter.

Like all software, BOINC has "features" that I don't like. As in all software use, well, I live with them. As an example, I never saw a stable version of Microsoft Word in my life (granted this was with 100+ page documents - MS help under pressure admitted that the intent was for ~25 pages, but I digress). However, we do have the chance to submit input. We can even submit code and have a decent chance to have the code wind up in the baseline.

That being said. The title of this thread suggests "Major Fixing" is required. As best as I can tell, there are two issues occurring with his situation. One is a network problem and the other is a mislike of the design of the configuration settings. As we did not get a request for help, nor a response I am not sure what is going on with the network. But, unlike the original SETI@Home Classic the BOINC networking is more complex to accomidate improved security and to eliminate the reliance on third party tools for things like proxies. With this can come problems. And some of them can be solved reasonably easy, others are much more difficult and take more time.

The configuration design is a little more difficult because it is the current design. Heck, in Classic you had no configuration at all. Now we have some. Yes, local settings would be nice. But, we don't have them yet. Maybe tomorrow, maybe next year, maybe never ... but even this could have been solved with some willingness.

Anyway, I don't want anyone to leave, neither do any of the other people whom I suspect you did not like their posts. But, they were not flames. Just simple factual statements, like: "you did not ask a question, ..."

We are all volunteers, those of us that try to answer questions. And those that are running BOINC (or trying to run BOINC). But, that is the point. We are volunteers, even Johnny. When you volunteer you do not get the right to grab the steering wheel. When you volunteer you live with situation that exists when you sign up ... When I volunteered for the US Navy I got the opportunity to live in conditions that judges were rulling unconstitutional for prisons (they did improve with time and promotions). But, I volunteered, so did you, so did he ...

If these are the biggest problems in his life then he should be truly happy ...
ID: 2207 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 2208 - Posted: 24 Dec 2005, 11:57:32 UTC - in response to Message 2205.  

I have several Macs that do not natively support OS X. Thus, moving them to BOINC has forced me to obtain, install, and configure a third-party patch to enable OS X installation.

Hi David,

I singled out that bit of your post, with a request. You could be of help, by testing the software. These instructions tell you how you can become a BOINC Alpha tester, if you want to do so. It's very difficult to release the software on a system with an OS when there were no testers for that system/OS.

The development team is looking for more people to test it on the Mac. As far as I know there are at present only a handful of people doing so. The multitude is testing on PC/Windows/Linux.

So you can help. It's just the question if you want to?

ID: 2208 · Report as offensive
Profile Tigers Dave

Send message
Joined: 24 Dec 05
Posts: 52
United States
Message 2215 - Posted: 24 Dec 2005, 20:38:58 UTC

Hi Jord:

I have signed up for the boinc_alpha mailing list and I will be alpha testing BOINC for Mac OS/X. Thanks for the information. Too bad that there are no plans to develop BOINC or BOINC clients for OS 9. Guess that I'll have to try to install LINUX/PPC on some of my old boxes; there is a BOINC client for that platform. Again, thanks.

David
ID: 2215 · Report as offensive
Profile Tigers Dave

Send message
Joined: 24 Dec 05
Posts: 52
United States
Message 2216 - Posted: 24 Dec 2005, 20:54:47 UTC - in response to Message 2207.  

We do try to be kind, generous and encouraging. And #4 *IS* the crux of the matter.

The title of this thread suggests "Major Fixing" is required. As best as I can tell, there are two issues occurring with his situation. One is a network problem and the other is a mislike of the design of the configuration settings. As we did not get a request for help, nor a response I am not sure what is going on with the network. But, unlike the original SETI@Home Classic the BOINC networking is more complex to accomidate improved security and to eliminate the reliance on third party tools for things like proxies. With this can come problems. And some of them can be solved reasonably easy, others are much more difficult and take more time.

Anyway, I don't want anyone to leave, neither do any of the other people whom I suspect you did not like their posts. But, they were not flames. Just simple factual statements, like: "you did not ask a question"


Hi Paul:

I agree with much of your assessment. I just wanted to remind the BOINC veterans that newbies like me and Johnny are not operating in a vacuum. We are posting our experiences and opinions to message boards other than this one. Indeed, even the SETI Classic message boards are still accepting posts. So, if newbies have bad experiences with BOINC and BOINC veterans, then they are going to flame BOINC and BOINC veterans in these other venues. And this bad publicity will keep others (particularly ex-SETI Classic crunchers) from joining BOINC. The trashing that BOINC was taking on the SETI Classic message boards was certainly a factor in delaying my transition to BOINC. So even when newbies act like jerks, I am asking the BOINC veterans to cut them a little slack and treat them with a respect that they may not deserve. In the long run, that will get more people excited about BOINC and distributed computing projects and will speed our progress toward the goals of these projects.

On a more personal note, I want to thank you Paul and Michael for all of the advice and assistance that you have given newbies like me. You guys probably don't get nearly the thanks that you deserve. Keep up the good work.

Best wishes and happy holidays.

David
ID: 2216 · Report as offensive
Paul D. Buck

Send message
Joined: 29 Aug 05
Posts: 225
Message 2223 - Posted: 25 Dec 2005, 3:45:53 UTC - in response to Message 2216.  

I just wanted to remind the BOINC veterans that newbies like me and Johnny are not operating in a vacuum. We are posting our experiences and opinions to message boards other than this one. Indeed, even the SETI Classic message boards are still accepting posts.

We know.

There has been a group of people that have dedicated quite a bit of time to the effort to bad mouth BOINC. BOth on the SETI@Home Classic site, and, on the SETI@Home Powered By BOINC site. Attempts to discuss matters with them usually prove to be fruitless. Many times when you boil down the argument it just come out that BOINC is not SETI@Home Classic, and they refuse to see the need for a change.

But, just around the corner is the first reason that BOINC is needed. The new enhanced application is shortly to be the new standard version. Following that are the AstroPulse and other efforts that will soon follow after that (we hope). The argument that Classic was flawless, and indeed from the participant prospective it did seem to be a very smoothly running system. Yet, I remember when I was running it that many times I would have to spend hours "baby-sitting" to fill up and empty my caches.

I also recall that the installation of my first CLI version did not go all that smoothly either. But, nostalgia aside, the architecture was fatally compromised and when it is that bad, starting with a clean sheet of paper is the best plan. Though many would (and don't like me to say this), but I do have system design experience ... and, in *MY* opinion, things are about as good as we have a right to expect. Not perfect, and in need of improvements, but, I have been running BOINC for nearly 2.5 years now and in almost all that time there has not been a day where I did not have work. Of course, I also listen to the man when he says that SETI@Home may not have work at all times and therefore it is a good idea to have other projects.

Anyway, common sense not being that common, we get people insisting that their "vision" is the only one that makes sense and if we would only listen to them, well, nirvanna ... Yet, when you try to pin them down on specifics you get hand waving. And loud exclamations of how good it will be, and oh, all the volunteers should be fired. This, usually because they cannot grasp that BOINC is very much a tiny project with very little budget. Along with that, so is SETI@Home. Even so, for a shoe string operation it works pretty well.

I do agree that is is possible that a abrupt answer may "drive" someone away. Well, when *I* look at the starting "questions" the person is not looking for help, they are looking for an excuse ...

Does that mean I think all the answers are perfect, nope, not mine, not anyone else's either. But, a person that asks almost always gets as much help as they can stand. I don't know, maybe I am just depressed by it all. I mean, its like I get requests all the time to be added to the editor list for the Wiki, yet, when asked what they would change, or to submit an article, silence ... same thing here ... BOINC stinks and don't confuse me with facts ... BOINC won't install but I don't want to tell you anything about my systems, or ry your suggestions and it is your fault because you did not fix my problem ... it does get old after awhile ...

On a more personal note, I want to thank you Paul and Michael for all of the advice and assistance that you have given newbies like me. You guys probably don't get nearly the thanks that you deserve. Keep up the good work.

You, and everyone else is welcome ...

Best wishes and happy holidays.

And the same for you.
ID: 2223 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 2228 - Posted: 25 Dec 2005, 10:49:11 UTC - in response to Message 2215.  

Hi Jord:

I have signed up for the boinc_alpha mailing list and I will be alpha testing BOINC for Mac OS/X.

Good to read that David. Make sure you bookmark the alternative link then. The Mac version at test is 5.2.15 for this moment. OS X again, I'm afraid.

Although you can always look around on the Other sources for BOINC Software link, try to contact people there to see if someone is attempting a port over to OS 9.

Then again, they didn't make the Windows version for 3.11 and only with limited success to work on Win95, so maybe it was just too anchient. ;)
ID: 2228 · Report as offensive

Message boards : BOINC Manager : BOINC needs some major fixing

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.