Boinc officially DOS'd my network

Message boards : Number crunching : Boinc officially DOS'd my network
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next

AuthorMessage
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225552 - Posted: 4 Jan 2006, 4:02:05 UTC - in response to Message 225544.  

I like normal mode and error mode.

Yes, we're well aware of that.

The goal is to never report work late. You'd be more unhappy if your results were routinely discarded because they were late.

Simple suggestion: set "connect every 'x' days" to one.

Please answer relevant question and not continue off topic. People at boinc don't like the truth. If the truth offends somebody let's call it something else.

So How do you cause only one computer to communicate with main servers and not cause DOS?

You've been complaining about Earliest-Deadline-First mode. I quoted your statement on that topic. Setting the connect interval to 1 day avoids EDF.

No I have tried to avoid responding to everyone who has brought it into this topic. See the big black letters at top. Can you read it?

ID: 225552 · Report as offensive
Profile KWSN - MajorKong
Volunteer tester
Avatar

Send message
Joined: 5 Jan 00
Posts: 2892
Credit: 1,499,890
RAC: 0
United States
Message 225561 - Posted: 4 Jan 2006, 4:10:34 UTC - in response to Message 225508.  



I like normal mode and error mode.
but that is not what this thread is for. Do you know how I can make it so only one computer caches work units for other computers and then that one computer returns the results.


You don't. This functionality has been discussed by the project staff, and some initial brainstorming was done, but they have other pressing matters that demand their immediate attention. It may get added at some indefinite time in the future, it may not. A lot depends on whether or not a.) a proposed design can be agreed on (by the project staff) that addresses the security concerns that prompted the initial design decision that each computer must communicate with the project(s) server(s) on their own to fetch and return work, b.) someone experienced with the BOINC code has the time, ability, and desire to code it up, and c) the code meets the approval of the staff for inclusion in the BOINC code-base. BOINC will not do this now, nor is it likely to do so for quite some time (if ever). If lack of this function hampers your participation in BOINC projects, I am sorry. There is nothing that almost everyone that posts here can do about it. The few posters that *can* are tied up working on other things at this time. Perhaps you could make more productive use of your time participating in another project (d-net comes to mind, with its 'personal proxy').

I guess the other question is why Boinc is not working correctly when trying to upload all the results at once. It goes only two deep the first two upload and go past 100% and then the next one starts while the other ones are still going.

I've never seen it work on more than 2 at a time, and after extended outages, I usually have a bunch to go back. And, believe me, I watch this closely.

The '>100%' issue on uploads is a known bug. It is being worked on.

From another post of yours:
An apology would be nice. Maybe a little help would be even nicer.


People have tried to help you. You, simply, just choose not to heed what they tell you. You just keep on whineing about whatever it is that bothers you today. For instance, take this thread. You described your problem when you started this thread (though you WERE more than a little trollish with your thread title (BOINC officially DOS'd my network)). As people chimed in, you rejected all their suggestions. You even admitted your router is sub-standard
The router doesn't switch lines too well. So if one computer is doing something it affects the other computers eventually.

So, somehow, in your tortured logic, you maintain that your difficulties are still 'all BOINC's fault', even though you have crappy equipment?

Nah, you are a troll. And you will get bugger-all of an apology from me.
https://youtu.be/iY57ErBkFFE

#Texit

Don't blame me, I voted for Johnson(L) in 2016.

Truth is dangerous... especially when it challenges those in power.
ID: 225561 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 225567 - Posted: 4 Jan 2006, 4:16:30 UTC

Hey Major.....who you talkin too?? Troll ignore list is activated!




Boinc....Boinc....Boinc....Boinc....
ID: 225567 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225574 - Posted: 4 Jan 2006, 4:24:31 UTC - in response to Message 225561.  



I like normal mode and error mode.
but that is not what this thread is for. Do you know how I can make it so only one computer caches work units for other computers and then that one computer returns the results.


You don't. This functionality has been discussed by the project staff, and some initial brainstorming was done, but they have other pressing matters that demand their immediate attention. It may get added at some indefinite time in the future, it may not. A lot depends on whether or not a.) a proposed design can be agreed on (by the project staff) that addresses the security concerns that prompted the initial design decision that each computer must communicate with the project(s) server(s) on their own to fetch and return work, b.) someone experienced with the BOINC code has the time, ability, and desire to code it up, and c) the code meets the approval of the staff for inclusion in the BOINC code-base. BOINC will not do this now, nor is it likely to do so for quite some time (if ever). If lack of this function hampers your participation in BOINC projects, I am sorry. There is nothing that almost everyone that posts here can do about it. The few posters that *can* are tied up working on other things at this time. Perhaps you could make more productive use of your time participating in another project (d-net comes to mind, with its 'personal proxy').


Supposed to be some way to set up one computer as a proxy.


I guess the other question is why Boinc is not working correctly when trying to upload all the results at once. It goes only two deep the first two upload and go past 100% and then the next one starts while the other ones are still going.

I've never seen it work on more than 2 at a time, and after extended outages, I usually have a bunch to go back. And, believe me, I watch this closely.

The '>100%' issue on uploads is a known bug. It is being worked on.

From another post of yours:
An apology would be nice. Maybe a little help would be even nicer.


People have tried to help you. You, simply, just choose not to heed what they tell you. You just keep on whineing about whatever it is that bothers you today. For instance, take this thread. You described your problem when you started this thread (though you WERE more than a little trollish with your thread title (BOINC officially DOS'd my network)). As people chimed in, you rejected all their suggestions. You even admitted your router is sub-standard
The router doesn't switch lines too well. So if one computer is doing something it affects the other computers eventually.

So, somehow, in your tortured logic, you maintain that your difficulties are still 'all BOINC's fault', even though you have crappy equipment?

Nah, you are a troll. And you will get bugger-all of an apology from me.


Let's see I couldn't connect with anything outside of my network. It was BOINC's Fault. US Robotic is supposed to be pretty decent. Most of the things said conflict with offcial design topics.

I have to bring his up: John told me that the problem with the scheduler is I connected to too many projects. I should just connect to one project. Scheduling for one project that's robust.

Basically everyone says except the crap you can't change anything. If anybody says anything that is anyway against your little world of BOINC you attack them. You can't say anything reasonable. You are keeping this topic open. I am not the only one who brings up the edf issue. This topic is too maybe get my computers back to crunching setiathome. You don't care about that. You don't want to do any real work and make the scheduler work correctly. What I want is what you should want it too.

How about answer a couple questions. How do you get the Boinc managers to talk to each other. How do you set up one boinc manager to act as proxy for the others. You are the one who sneaks around and attacks everyone.

ID: 225574 · Report as offensive
Profile Siran d'Vel'nahr
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 7379
Credit: 44,181,323
RAC: 238
United States
Message 225576 - Posted: 4 Jan 2006, 4:24:49 UTC - in response to Message 225471.  

....
....
....
You are too happy to support substandard program.

I see NOTHING "substandard" about BOINC....
All I have been doing is install connect and let it run.

YEP! That's right, install it, start it and let it run. Who knows better how BOINC is to run than the devs? You know? As Tony mentioned in a previous post, we have gone over this in at least 2 other threads. You attitude that "I'm right and everyone of you are wrong" is not going to get you any sympathy, let alone help.
I have been running boinc for a couple years.

I don't believe the above statement for a minute. About a year ago, I read somewhere here that there was some BOINC bashing going on over at Classic. I went and checked it out for myself. I found that it was you and several others doing the bashing. You were NOT doing BOINC at that time.

Let's look at your credits and mine:

I started doing BOINC on one PC. I added a second several months later, a lesser PC than my XP. I decided to switch OSs on the second PC to a non-Winblows OS. For another few months, that PC did not crunch. About 2 months ago, that PC started crunching.

You say you have a network of 5 PCs crunching. IF you have been doing BOINC for "a couple years", why is it that my credits, with one full time PC and one part time PC crunching, are far and away greater than yours? I'm over 80K, you're just over 12K. Care to rescind you statement? Or, give a believable explanation as to why your statement is "true" in your eyes? If I'm not mistaken, I believe it was early 2005 when you started doing BOINC. (-:<

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath
ID: 225576 · Report as offensive
Profile Tern
Volunteer tester
Avatar

Send message
Joined: 4 Dec 03
Posts: 1122
Credit: 13,376,822
RAC: 44
United States
Message 225584 - Posted: 4 Jan 2006, 4:39:01 UTC - in response to Message 225576.  

If I'm not mistaken, I believe it was early 2005 when you started doing BOINC. (-:<


The troll's first BOINC computer was attached on 20 Apr 2005 13:47:38 (a Celeron). His second was 16 Dec 2005 1:23:34, and his latest (fifth) was 17 Dec 2005 22:03:52.

He obviously set up one system early in the year, didn't like what he saw here and was unable to understand it, and went back to Classic where he could piss and moan about how lousy BOINC was, until Classic finally shut down. Now he just wants to do as much damage as he can before he gets banned from the boards.
ID: 225584 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225585 - Posted: 4 Jan 2006, 4:39:35 UTC - in response to Message 225576.  

....
....
....
You are too happy to support substandard program.

I see NOTHING "substandard" about BOINC....
All I have been doing is install connect and let it run.

YEP! That's right, install it, start it and let it run. Who knows better how BOINC is to run than the devs? You know? As Tony mentioned in a previous post, we have gone over this in at least 2 other threads. You attitude that "I'm right and everyone of you are wrong" is not going to get you any sympathy, let alone help.
---rest cut----


I don't want to see or have scheduler in computer is overcommitted when it is not. The scheduler is not working properly else I would have 1 CPDN, 3 Einsteins, and 3 seti@homes on the one computer. I would have CPDN run a couple hours, Einstein would run a couple hours and seti@home would run an hour. If one project would somehow go late, give a little time. Not all of the time.
I get 150K/b download average. I take advice that doesn't go against what boinc was designed for.

ID: 225585 · Report as offensive
John McLeod VII
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jul 99
Posts: 24806
Credit: 790,712
RAC: 0
United States
Message 225591 - Posted: 4 Jan 2006, 4:45:33 UTC - in response to Message 225574.  

I have to bring his up: John told me that the problem with the scheduler is I connected to too many projects. I should just connect to one project. Scheduling for one project that's robust.

If you INSIST on running so many projects with such a huge cache that the scheduler believes that it needs to keep the host in EDF, you should not complain about it. If you want to avoid EDF, reduce your cache size, or reduce the number of projects. Large caches are needed for those running normally disconnected, or with very few projects. Large caches are NOT needed by those running more than 2 projects that are normally connected.

How about answer a couple questions. How do you get the Boinc managers to talk to each other. How do you set up one boinc manager to act as proxy for the others. You are the one who sneaks around and attacks everyone.

Use freeproxy or similar.


BOINC WIKI
ID: 225591 · Report as offensive
Profile Siran d'Vel'nahr
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 7379
Credit: 44,181,323
RAC: 238
United States
Message 225592 - Posted: 4 Jan 2006, 4:45:45 UTC - in response to Message 225508.  

....
....
I like normal mode and error mode.
but that is not what this thread is for. Do you know how I can make it so only one computer caches work units for other computers and then that one computer returns the results. I guess the other question is why Boinc is not working correctly when trying to upload all the results at once. It goes only two deep the first two upload and go past 100% and then the next one starts while the other ones are still going.

Has it occurred to you, yet, that you may just have been given the correct answer and you refused to see it? You say you have a connection problem when more than one of your PCs try to connect at the same time. That should give you a clue. My 2 PCs sit behind my router connected to my cable modem. They have no problem access the Internet at the same time. I have downloaded files on one while posting here, there, wherever on the other at the same time. I have seen both PCs accessing SETI at the same time up/downloading WUs. I would say that there is a problem with you connection to the Internet or that you have a network problem.

So, instead of spouting off here about how you're right and everyone else is wrong, why don't you look at things on your end and figure out YOUR problem. It is NOT a BOINC problem. (-:<

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath
ID: 225592 · Report as offensive
John McLeod VII
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jul 99
Posts: 24806
Credit: 790,712
RAC: 0
United States
Message 225595 - Posted: 4 Jan 2006, 4:50:35 UTC

I had a USR router at one point. It was pure crap - several of the functions that it was supposed to support did not work correctly, and one of them was critical to me. I ended up with a different brand of router, and the problems were fixed.


BOINC WIKI
ID: 225595 · Report as offensive
Profile Pappa
Volunteer tester
Avatar

Send message
Joined: 9 Jan 00
Posts: 2562
Credit: 12,301,681
RAC: 0
United States
Message 225596 - Posted: 4 Jan 2006, 4:53:53 UTC
Last modified: 4 Jan 2006, 4:56:15 UTC

Can we please have a Moderator Step In... Everyone has lost track of what might really be an "Issue," and everyone else thinks what they thought what they read is not really what is happening... There is really nothing "productive" here...

Most new people come in with "Issues" Many here help as best they can... Then everyone becomes happy... This particular thread has gone from bad to good, then back to bad again...

What is wrong with this picture? Silly me am here to do the Science, and help as best I can...

R/

Al

Please consider a Donation to the Seti Project.

ID: 225596 · Report as offensive
Profile Siran d'Vel'nahr
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 7379
Credit: 44,181,323
RAC: 238
United States
Message 225601 - Posted: 4 Jan 2006, 4:58:57 UTC - in response to Message 225584.  

If I'm not mistaken, I believe it was early 2005 when you started doing BOINC. (-:<


The troll's first BOINC computer was attached on 20 Apr 2005 13:47:38 (a Celeron). His second was 16 Dec 2005 1:23:34, and his latest (fifth) was 17 Dec 2005 22:03:52.

He obviously set up one system early in the year, didn't like what he saw here and was unable to understand it, and went back to Classic where he could piss and moan about how lousy BOINC was, until Classic finally shut down. Now he just wants to do as much damage as he can before he gets banned from the boards.

Thanks Bill, I knew I was right in that he hasn't done BOINC for a couple years and just started early last year. I love catching trolls in a lie.... (-:<

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath
ID: 225601 · Report as offensive
Profile Siran d'Vel'nahr
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 7379
Credit: 44,181,323
RAC: 238
United States
Message 225620 - Posted: 4 Jan 2006, 5:14:35 UTC - in response to Message 225585.  

....
....
....
....
....
....
---rest cut----
....

Ok troll, since you won't listen to what people are telling you check this out:

I see that you have skirted around the LIE I caught you in. Come on, what's up with that? You are a LIAR, ksnash! I'll bet that every one of your past threads, and this one, are all lies. You have been at 12K credits for quite some time now. I saw you were at 12K credits a month or so ago when you started your BOINC bashing rampage back then! Now you continue with your BOINC bashing rampage here. You most likely started your BOINC bashing rampage when you were still a noob on Classic. You're still a noob, you will always be a noob. As long as you refuse to listen to what people tell you here, you will remain a noob. You are a troll and you are a noob, but, you are not a noob troll because you have trolled before many times and many months ago. If you HATE BOINC THAT MUCH why don't you just uninstall it from your network of PCs and go play with yourself for the rest of time.

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath
ID: 225620 · Report as offensive
Profile KWSN - MajorKong
Volunteer tester
Avatar

Send message
Joined: 5 Jan 00
Posts: 2892
Credit: 1,499,890
RAC: 0
United States
Message 225652 - Posted: 4 Jan 2006, 6:12:39 UTC - in response to Message 225574.  
Last modified: 4 Jan 2006, 6:13:39 UTC



Let's see I couldn't connect with anything outside of my network. It was BOINC's Fault. US Robotic is supposed to be pretty decent. Most of the things said conflict with offcial design topics.

The cause of your up/download issue is most likely your router. Try a different one. Borrow one from a friend, if you have to. US Robotics may (or may not) be a decent brand... The model # you have may, in most all cases, work perfectly. But I think that the cause of the up/download issue is that your SPECIFIC router is flakey. Try a different one, and see if it helps.

I have to bring his up: John told me that the problem with the scheduler is I connected to too many projects. I should just connect to one project. Scheduling for one project that's robust.

What he said was
If you INSIST on running so many projects with such a huge cache that the scheduler believes that it needs to keep the host in EDF, you should not complain about it. If you want to avoid EDF, reduce your cache size, or reduce the number of projects. Large caches are needed for those running normally disconnected, or with very few projects. Large caches are NOT needed by those running more than 2 projects that are normally connected.


What he said was to either reduce the number of projects OR reduce your cache size if you want to avoid EDF mode.


Basically everyone says except the crap you can't change anything. If anybody says anything that is anyway against your little world of BOINC you attack them. You can't say anything reasonable. You are keeping this topic open. I am not the only one who brings up the edf issue. This topic is too maybe get my computers back to crunching setiathome. You don't care about that. You don't want to do any real work and make the scheduler work correctly. What I want is what you should want it too.


One should never just blindly accept crap. That is why a lot of people respond to you negatively; we don't accept YOUR crap. I am not a blind cheerleader for BOINC. I am one of the first to say that BOINC still has a few warts. That said, there is a proper way to submit bug-reports for actual BUGS (ever hear of BOINCZilla? I thought not). Most actual bugs get fixed rather quickly. Out of several bugs that I have reported, I have only had one BUG-report that has not been fixed yet. They thought they had it fixed, as did I, but it has recently resurfaced. I will re-submit it once the dust from the closure of Classic and the ramp-up of _Enhanced has settled. Now, this won't work if you are not reporting a bug, but instead have an issue with a design-decision. To get one of the staff's design-decisions changed takes an Act of God and Congress. And repeatedly making a pest of yourself here on Number Crunching by whineing about how one of the design-decisions is a royal pain in your backside will NOT get anything done about it.

I didn't write the scheduler. Besides, it works correctly for me and a great many other people. If you want to get your computers back to crunching S@H, please follow the advice you have been given. Try another router. If you want to avoid EDF-mode, reduce your cache size, or number of projects on each host.


How about answer a couple questions. How do you get the Boinc managers to talk to each other. How do you set up one boinc manager to act as proxy for the others. You are the one who sneaks around and attacks everyone.

Ok...

#1: You don't. You CAN get one BOINCManager to talk to several BOINC clients, but this won't help you. The BOINC client is what does the uploads and downloads, not the BOINCManager. All BOINCManager is is a GUI control panel for one (or more) BOINC client(s).

#2: You don't. Neither BOINCManager nor the BOINC client has proxy capability built in. Now JM7 mentioned freeproxy, which is a standalone proxy program you can install on one of your boxes. Now, there are circumstances where this would be helpful, such as if only some (or one) computers on a LAN could 'reach' the internet due to usage/security restrictions. Put freeproxy on one of the computers that CAN talk on the internet, and point the BOINC clients on the ones that can't to that proxy. But in your case, I think this would be counterproductive. It would effectively double the traffic thru your router/switch. Instead of "Computer->router/switch->internet" it would be "Computer->router/switch->computer with proxy->router/switch->internet".

Take my advice: try another router, and see if that fixes things.

I don't "sneak around", and I only flame trolls that deserve it.


https://youtu.be/iY57ErBkFFE

#Texit

Don't blame me, I voted for Johnson(L) in 2016.

Truth is dangerous... especially when it challenges those in power.
ID: 225652 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 225653 - Posted: 4 Jan 2006, 6:16:51 UTC - in response to Message 225552.  


No I have tried to avoid responding to everyone who has brought it into this topic. See the big black letters at top. Can you read it?

Okay fine.

Routers switch packets, they don't switch wires.

If one of your computers is uploading to Berkeley, it gets all the bandwidth.

If two are uploading, they'll get about half.

If all five are uploading, then you divide the bandwidth five ways.

This isn't a strict division: packets come from all five computers to the router. The router sends them out. Each computer will stop sending when a certain number of bytes. According to this knowledgebase article it's something like 8k. The remote sends an ACK and things continue.

I don't have a lot of results on my machine at the moment, the one I have in progress is 10k bytes, or 80k bits. Add 20% or so for overhead. At 256kbits/sec., let's round up to 1/2 second per result unit.

If your five machines each have 20 results to upload, that's 100 work units total, or about fifty seconds of your full bandwidth. Round up to a minute just to make the math easier.

Now, that's in a fairly perfect world. Any bottlenecks between you and Berkeley will slow things down, but that also means your bandwidth will be underutilized.

If we just plain assume that I'm off by an order of magnitude, then yeah, I can see how your network was running at capacity for as much as ten minutes.

If it was longer than ten minutes, you need to look at other issues: bad router or network card, inappropriate TCP parameters, or some other issue not related directly to BOINC.
ID: 225653 · Report as offensive
Profile KWSN - MajorKong
Volunteer tester
Avatar

Send message
Joined: 5 Jan 00
Posts: 2892
Credit: 1,499,890
RAC: 0
United States
Message 225655 - Posted: 4 Jan 2006, 6:18:29 UTC - in response to Message 225567.  

Hey Major.....who you talkin too?? Troll ignore list is activated!





Yeah, I *COULD* stick him on ignore... But where is the fun in that? <evil_grin/>
ID: 225655 · Report as offensive
Mark Day
Avatar

Send message
Joined: 19 Aug 02
Posts: 81
Credit: 502,830
RAC: 0
United States
Message 225703 - Posted: 4 Jan 2006, 9:55:34 UTC
Last modified: 4 Jan 2006, 10:00:46 UTC

Love the topic, but the obvious answer is... get a better network. If the OP's problem was not "unique" to his setup, who here would doubt that there would be hundreds of "me too's" chiming in. Notice their conspicuous absence.

M
ID: 225703 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13715
Credit: 208,696,464
RAC: 304
Australia
Message 225721 - Posted: 4 Jan 2006, 11:07:17 UTC - in response to Message 225519.  

An apology would be nice.

We're sorry you're having trouble coping with reality.


Maybe a little help would be even nicer.

Your only purpose in posting is to whinge & whine. You ignore any advise you're given, so why ask?

Do everyone a favour & either use the advice that people give you, or just don't bother posting in the first place.
Grant
Darwin NT
ID: 225721 · Report as offensive
Astro
Volunteer tester
Avatar

Send message
Joined: 16 Apr 02
Posts: 8026
Credit: 600,015
RAC: 0
Message 225741 - Posted: 4 Jan 2006, 12:32:42 UTC - in response to Message 225508.  

[ It goes only two deep the first two upload and go past 100% and then the next one starts while the other ones are still going.

See Boinczilla bug report #366, It's already been reported.
ID: 225741 · Report as offensive
Jack Gulley

Send message
Joined: 4 Mar 03
Posts: 423
Credit: 526,566
RAC: 0
United States
Message 225758 - Posted: 4 Jan 2006, 13:30:05 UTC - in response to Message 225703.  

Love the topic, but the obvious answer is... get a better network. If the OP's problem was not "unique" to his setup, who here would doubt that there would be hundreds of "me too's" chiming in. Notice their conspicuous absence.

M

A lot of other people are having the same sort of problems. See the same things they don't like about BOINC. And have little confidence in the BOINC program right now. ksnash just does such a good job keeping things stirred up that the rest of us don't have to chime in. We can just set back and read and follow along.

Actually, there has been some real good useful information posted in this thread trying to shut him up. (Thanks KWSN.) Probably help more than a few people "workaround" their problems with BOINC.

I think ksnash is like most of us "pre NEW MATH" people who look at something that sets there and says, Progress 100.00% uploading, and assume that means what it says, that either it is finished or it is proceeding a 100% of the bandwidth of the network. Not that it is just setting there doing nothing. And as for 108.71%, well ..... Does not inspire confidence that the program is working correctly. I sure hope those changes and fixes start off with the Progress setting at 0.00% long before it attempts to connect instead of setting there showing 100.00% before it even starts. Would be nice if the program sets it to 0.01% when the connection is started. And then updates the Progress after a block of data is sent and acknowledged, not when the programmer sends it to the output buffer to hopefully be sent. It does not inspire confidence in a program that is telling you its Progress is at 8.69% when any fool with a packet trace program can see that not only has it not sent anything, the server told it not to send anything. Yet the program sets there timing out, showing an invalid progress.

Little things like that, and error messages that do not mean what they say do not inspire confidence that a program is working correctly. And with out that confidence many people are not going to just set back, trust it, and let it run, without checking on it frequently, and punching buttons.

You would think that the users who have been using and testing BOINC for the past year or more, would have long ago noticed such problems and saw to it they were fixed.

When someone sees an error message that relate to problems they have, they assume that program is the cause of all their related problems.

ksnash does need to either get a better router or setup up his router firewall better. I had problems for a long time with the default MultiCast Passthrough enabled and other default PassThrough settings for a long time. Found about a Mbps of this junk coming into my local network most of the time, slowing other traffic down and actually effecting the systems having to receive and reject all of it. One would do randome reset when trying to access web pages. Seems I am now on an ISP's subnet with a major collage many miles away, and they have some suspect routers broadcasting traffic to everyone on that same ISP subnet. At least now my router is set to block all of it. ksnash needs to check his router settings to do the same, and not assume the default PassThrough setting are OK.

As for overcommitted problems, there are some problems there. I saw one I could not explain, but did not matter as BOINC was to be removed from it later that day. I had three machines get into that state with only a 1 day queue setting. Could not get it to request work, until it ran out of work and set there for 10 minutes doing nothing. Then I could manually get it to download one or two WU. Took three days to get it sorted out by setting the days to 10, and punching buttons every few hours to get to to download work when it ran out.

ksnash, by chance is the machine(s) that is overcommitted one that you had BOINC running on while you still had Seti Classic running on it. If so, all the projects you were running at that time are now messed up on how long it takes to do a WU. Check and see if the estimated times are several times longer than it actually takes. It looks at these estimates, not the actual times to determine if it is overcommitted. There is no user interface to fix it. What you will have to do is set "No new work" for all of your projects and let all of your current work run out. Then enable work for one project at time, say Seti first, and then set the days to 10 and keep checking it and downloading new work for several days, until the estimated time for the ready to run UW drop to only about 50% more than they actually take. Then cut the days back down to 1 once you have enough work to keep it busy for a full day and run it until the estimated times are close to the actual time it takes. Then shut that projects new work off and run the queue down, and go through the same procedure for each of your projects that have high estimates. Once you have done enough work that the estimates for each project are close to accurate, you can enable all of them again, and things will start working a lot better. Just don't use the machine for other things that might cause long run times for the active WU's.

Would be nice if there was some sort of interface that allowed users to reset or fast adjust this setting when it gets messed up. But then BOINC is not about user interfaces.

I have two older slow machines I can not connect directly to the Internet for security reasons, and do not have a backup for the others when the cable goes out for extended periods, and need a program like SetiQueue to handle it. With SetiQue, I could use a dial-up for backup. A few times i just moved the SetiQue system to a location where there was a good Internet connection, to get/send work for the other machines. A proxy will only solves part of the problems.

There are others that need to chime in now and then to remind all, its not just a few people with problems.
ID: 225758 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next

Message boards : Number crunching : Boinc officially DOS'd my network


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.