Shutting Down the "Big Rigs"

Message boards : Number crunching : Shutting Down the "Big Rigs"
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3

AuthorMessage
Bree

Send message
Joined: 7 Feb 00
Posts: 9
Credit: 12,468,505
RAC: 0
United States
Message 1178396 - Posted: 16 Dec 2011, 2:53:21 UTC
Last modified: 16 Dec 2011, 2:59:59 UTC

Well got up to 399 wu's then <sarcasm> couldnt up/down load for 6 hours </sarcasm> now it working again but ONLY thru a proxie. in every 10 D/L I get 4 that have to retry up to 4 times <and 4 to 6 hours backoff total time> so i get 10 in about 6 hours.. this is BAD HORRIBLE UNACCEPTABLE!!

I look at it this way.. Seti wants me to give. I am.. My time , computer, and power. I am happy to do so .. and happy to help more IF i got reliable connections.

Now a lot get mad at WU crunchers .. Well to bad most of us crunch and I do
1. for the Joy of helping maybe finding ET's or something New out here.
2. for the Pride in helping
3. for the Place in the rankings that shows 'I' am helping science.
4. for the Honor in what I have done , ranking.

so for the Joy, Pride, Place, and Honor.. This is why I crunch and why I get upset that I get very little feedback, poor support in both program and science, lack of feeling wanted in the help I give. Then get asked for more, why?

It's not the big guys only getting let down. Its the small crunchers like me. So it seems the politics and authorities at school, routing, ISP's and in the science tend to hurt this.

This morning I had 0-ZERO work for 4 cores, 1 GPU and hung uploads. as of right now, I have 312 gup 172 cpu and 2 astro's and 1 hung down load with 2+hour retry time. I have had to retry manually all the stuck up/downloads today STILL. total over 80.

So should I leave seti till they fix it <and I love the idea of seti as to me its the one thing I really care about>. Should I manually have to retry all the hangs , which are a pain but clog the system so bad i cant keep crunching. Or do I and others start DEMANDING as the providers of the support for the science, MORE communications and fix for the problem before we give more.

I know a lot of people think about leaving or staying. I want to stay , I want to feel useful to the project, but the treatment from SETI COMMAND is almost an afterthought for the users , both big farms and small guys. We got in this because we Dream there is ET somewhere. We helped because it is the dream were after. But without feeling needed by SETI COMMAND not other users , not second hand comments from others but SETI COMMAND itself, many are going to choose to leave. This is already happening. Without SETI COMMAND stepping up and taking the time to keep the user base informed they will loose the very support that makes the science possible.

6+ days on no info, no action on network problems other then a notice to say it was fixed and for me nothing changed or got worse. It used to be a regular weekly thing to get good updates.. it should be at least 2 to 3 times a week just a note to let us know what the holdups are or how the science is going. The vacuum around the project has got to end before it ends the great community of volunteers that help is some of the greatest science in the history of man.

Rig
AMD 955 OverClocked to AMD 975
4 Gig
GTX 280
DVD Burner
500G Raid Mirror
500G Main HD
WinXP/3
all water cooled <custom parts build>
and I run it 24/7 to crunch:) <anyone know where i can get a couple of cheap 1TB drives for the raid..>

Haf932
ID: 1178396 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 1178404 - Posted: 16 Dec 2011, 3:47:34 UTC

If an E7600 + GTX260 can't get enough S@h WUs for 24/7, I don't need to think about to switch on again my big machine (940 BE + 4x GTX260)..

Really sad.. - my small machine can reach the S@h server, but continuously backlogged DLs.. only sometimes only a few WUs come very slowly down..


- Best regards! - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC. - SETI@home needs your help. -
ID: 1178404 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1178443 - Posted: 16 Dec 2011, 9:22:06 UTC - in response to Message 1178396.  

Bree, my man.....
Don't know what I can say to you.
This IS a volunteer project, and anybody is as free as the breeze to come or go as they wish.

If ya wanna do Seti, you are just gonna have to roll with the punches.
Nobody more than I would love 24/7 uptime and unlimited bandwidth to support the ever increasing demands of faster and faster computers with faster and faster GPUs.

Nobody on this project bleeds Seti more than I do when things are not flowing well.
But, as far as the project goes, even though there are many times when the servers cannot fully support a fast multi GPU rig, the project itself is getting more work processed for it than ever. Up to the limits of it's hardware.

The GPUUG fundraising efforts are addressing the hardware needs, as requested by the staff. We are not just guessing what they want, they are telling us.

So....complaining that the current situation is 'bad horrid unacceptable' is just so much bluster.
Unless you are the one with a half a million bucks or so burning a hole in your pocket to make a dent in the utopian setup we all hope for.......
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1178443 · Report as offensive
Treasurer

Send message
Joined: 13 Dec 05
Posts: 109
Credit: 1,569,762
RAC: 0
Germany
Message 1178455 - Posted: 16 Dec 2011, 10:29:52 UTC

I wonder: does SETI upload the WUs faster then Arelcibo is recording(the tapes are "only" a couple of months behind)? If so, we dont need more bandwidth.
ID: 1178455 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1178459 - Posted: 16 Dec 2011, 10:43:31 UTC - in response to Message 1178455.  

I wonder: does SETI upload the WUs faster then Arelcibo is recording(the tapes are "only" a couple of months behind)? If so, we dont need more bandwidth.

Good question.......

I am not sure about the amount of datasets in backup......
But the among the other things that the GPUUG fundraising efforts is addressing is more 'shuttle drives' to get data to and fro from Arecibo.

The one thing we do not have any control over is the nature of the recordings we get from our time on the telescope.

We only 'piggyback' on whatever other projects are running on the t-scope at the time, and have no say in what scans are being conducted.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1178459 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1178499 - Posted: 16 Dec 2011, 14:32:18 UTC - in response to Message 1178396.  

Bree,

Here's how I look at it. If S@H can't give you any work to do for it, that is its problem, not yours. Ultimately, how does it hurt you if you can't get any work to do? It's not like you're getting paid to do it and losing income when you can't; just the opposite, it costs you money to do it and you save when you can't. Sure, the guys in the lab would like to be able to push through all the data they want without ever having a hangup, but their problems take money to fix and they don't have it, and in the end, it's only skin off the project's nose, not yours.

Also, the time they spend telling us about their problems is just time they can't spend working on those problems. Yeah, it's frustrating not knowing why most of the servers went out to dinner Wednesday night, but they fixed it first thing in the morning and then presumably went on with whatever else they had planned for the day. (That said, dammit Matt, can't you take 3 minutes to tell us what happened??? :-)

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1178499 · Report as offensive
Bree

Send message
Joined: 7 Feb 00
Posts: 9
Credit: 12,468,505
RAC: 0
United States
Message 1178839 - Posted: 17 Dec 2011, 13:42:58 UTC - in response to Message 1178499.  
Last modified: 17 Dec 2011, 13:45:03 UTC

The point is missed from my post.

I love what I do and enjoy Seti crunching. I don't love 11 hour back-offs for dropped wu. Its not about getting work all the time, been with Seti to long and KNOW they run out of work for many reasons, I can live with no work when it truly is 'we have no work'.
What I have a hard time understanding is why when i get WU's its not reliable. Having to babysit and hit retry to just get the DL unstuck. I just want the basics of boinc/seti to give reliable UP/DOWN loads IF work is available and YES, I DO get payed!!
I get payed in the satisfaction that I a little guy out in podunk nowhere can actually help in a dream. So when I don't get good loads<should I even say it that way?? :)> I get frustrated that its hung again, the dream is on hold.. and the big point, I AM part of the project so it IS skin off my nose that I cant crunch more for them. Its my passion to someday see Seti find anything useful and it already has done a lot.
I dont really care why the servers went down, to me its nice to know why everything hung, however what I am asking for is a fix for the UP/DOWN loads are hanging over the long haul. What is planed, or asked for, to help fix the problem and what will it do. We need more then a snip, like, we want drives because we need storage. Let us know what part of the science will be helped, what does it do to further Seti. Include us, the crunchers, this is what I am seeing fall away. The idea of your a volunteer and therefor shut up or leave if something is not the way you like it, is a bad precedent. So its my passion for the dream and for the science i get upset and down in the dumps when i dont get reliable UP and DOWN loads IF work is available and showing in the boinc client as UP/DOWN loading. Like I asked before WHAT do we <the volunteers> need to do to help get a better link so we can help Seti go more science.
This is not about baiting or as some say whining. Its the voice of one dreamer wanting to help crunch what he can AND communicate the work TO and FROM back to Seti in a reliable way. Not watch is puter sit idle as a WU is hung in a 6 8 15 hour wait because of cant connect to peer, ect, for the 9th retry causing the beast to run out of work when lots of work is available. [humm maybe get Goggle or Amazon to donate bandwidth?? ;) ]
Sorry to all, I do get very passionate on doing this work. I wont be going on about this as it is the frustration of poor up/downs that's got me going. Not not having work to do but the lack of functional working of the basic links.

Dream to Live, Live to Dare, DARE TO DREAM!
ID: 1178839 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1178843 - Posted: 17 Dec 2011, 13:53:03 UTC - in response to Message 1178839.  
Last modified: 17 Dec 2011, 13:53:56 UTC

Your point is well said and taken.

I agree that the Boinc setup is far too 'protective' of the project servers.
But it has to function for all Boinc projects, not just Seti.

For Seti, it is truly whacked. Asinine backoff schedule and project deferrals. Makes it almost impossible to achieve sustained contribution to the project without manual intervention. Makes requesting and downloading work the exception rather than the rule.

Just whacked.

Anybody have the tools, time, and patience to recomplile a proper Boinc for Seti?
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1178843 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 9786
Credit: 572,710,851
RAC: 3,799
Panama
Message 1178861 - Posted: 17 Dec 2011, 15:22:54 UTC
Last modified: 17 Dec 2011, 15:27:22 UTC

Hi

I have a sugestion to temporaly fix the BW problem, probabily looks like something stupid but works in the company i work for.

When we have some serius BW problem in our 100mbps line (of course less than Seti but something like that because we need to increase the BW and cant make the investiment to change optical fiber needed to increase the BW) we lease 2 new lines (50 mpbs each) for a very cheap price (even in our tax hungry country) for another internet company (one ADSL for local telefone company and One Cable from local cable TV) and split the trafic in the 3 lines.

Can someting like that could be done in Seti? At least until the 1GBps line realy reach Seti? This is not an elegant solution but probabily will give the time needed to make the long term fix of the BW problem and certainly give us the users less stress and retrys and of course the instalation cost of these simple line is very cheap and requires no construction or heavy changes.

Just a sugestion, i am not a computer/network expert, just from the user point.
ID: 1178861 · Report as offensive
Profile Donald L. Johnson
Avatar

Send message
Joined: 5 Aug 02
Posts: 8240
Credit: 14,654,533
RAC: 20
United States
Message 1178971 - Posted: 17 Dec 2011, 22:21:36 UTC - in response to Message 1178861.  

When we have some serius BW problem in our 100mbps line (of course less than Seti but something like that because we need to increase the BW and cant make the investiment to change optical fiber needed to increase the BW) we lease 2 new lines (50 mpbs each) for a very cheap price (even in our tax hungry country) for another internet company (one ADSL for local telefone company and One Cable from local cable TV) and split the trafic in the 3 lines.

Seti@Home is hosted by the University of California.
The University, not S@H, controls their Internet access.

Even if S@H had control, it would still cost money to lease additional access lines, money they don't have right now, unless/until WE donate it.

Donald
Infernal Optimist / Submariner, retired
ID: 1178971 · Report as offensive
Profile lupo

Send message
Joined: 29 Aug 10
Posts: 91
Credit: 4,736,407
RAC: 0
United States
Message 1179022 - Posted: 18 Dec 2011, 5:34:18 UTC

I'm in the same boat when it comes to now doing work for SETI any more. I already had 1 high end system dedicated to SETI processing and after another $7k investment I built a beast of a CUDA system specifically to help out with SETI. After I finally got the server built SETI really stared having problems. For the first 10 days of the system being built I couldn't download any WU from SETI! Now the issues with SETI have been dragging on for months and I STILL am unable to reliably download data from the SETI servers.

As someone that knows how to build and maintain high availability data centers I can't believe how shoddy of a setup the folks at SETI appear to be running. If I remember correctly, a few months ago there was a serious problem with the servers because they did not have something as simple as a stable battery backup!

So now, I have pretty much given up on SETI and have just been processing Einsein@home. I never seem to have any issues getting WU from their servers.:/

Honestly I don't think lack of funding is the only problem, I think it is also the people handling the architecture and current operation of the SETI servers.

Overall, I just feel sad about this situation. I'll keep SETI in my work queue as I crunch for Einstein and hopefully get some Wu here and there. Please get your act together SETI. Your work is very import ant and there are many of us who want to help. When after so many months you still are experiencing the same issues over and over it appears to people on the out side that you don't really care, therefore why should we all care about dedicating our resources to help.

ID: 1179022 · Report as offensive
Terror Australis
Volunteer tester

Send message
Joined: 14 Feb 04
Posts: 1817
Credit: 262,693,308
RAC: 44
Australia
Message 1179027 - Posted: 18 Dec 2011, 7:08:29 UTC

Well, the Big Rigs are back up and crunching happily, all boxes are on the limits and I'm smiling :-)

My only problem now is the continuing battle to find working proxies to get around the intermittant connection problems between Reach and Hurricane at the PAIX. I've decided to spread the load and have 4 boxes connecting through 4 different proxies, 2 in Germany, 1 in Brazil and 1 in the UK. This stops the Sysadmins freaking about the bandwidth usage and blocking the SAH ip's. I also have one box on direct to monitor the servers.

It's interesting. The direct download speeds vary from zilch to 20kBs while through the proxies it's constant at around 10kBs stopping only when when there's a total freeze up at the SAH servers.

T.A.
ID: 1179027 · Report as offensive
Profile Gatekeeper
Avatar

Send message
Joined: 14 Jul 04
Posts: 887
Credit: 176,479,616
RAC: 0
United States
Message 1179030 - Posted: 18 Dec 2011, 7:45:34 UTC
Last modified: 18 Dec 2011, 7:46:36 UTC

For those of you who've heard what I'm about to post many times over, my apologies. But I feel the need to again point out what I thought was obvious, yet, it keeps being repeated in various threads throughout these forums: that is, "why can I go to (insert project X here) and get all the work I want with no problems, but I can't at Seti?"

From the list of top BOINC projects:

    Collatz Active Users 6,634 Active Hosts 10,724
    Milky Way Active Users 26,423 Active Hosts 36,836
    GPUGrid Active Users 2,408 Active Hosts 3,174
    World Comm Grid Active Users 93,573 Active Hosts 216,561
    Einstein Active Users 40,043 Active Hosts 71,500
    Rosetta Active Users 37,401 Active Hosts 59,516
    Seti@Home Active Users 152,096 Active Hosts 235,318



With the exception of WCG, which is the best funded project by far on BOINC (take a look at their 45 pages of sponsors, including IBM), the total active users and hosts of all the top projects on BOINC is less than that of Seti alone. Even WCG isn't as large.

And Seti is pretty much an orphan project. If it weren't for the generosity of its users, the ability of the project to persevere would be questionable at best.

Now, there is no doubt that Seti appreciates each and every one of the volunteers who crunch for the project. But all the hours of crunching you do, and the expense you lay out for equipment and power, don't add a dime to the bottom line at Seti to address the problems that exist. If you want to vent about what you perceive to be shortcomings in the project, fine, that is your right. But try and look at things from the viewpoint that Eric, Matt, Jeff and the rest of the staff must have as they try to sustain the project on a shoestring budget AND advance the science. Maybe at that point, your perceptions might be slightly modified.

There, I've said it. Again. I feel better. Again.


ID: 1179030 · Report as offensive
Terror Australis
Volunteer tester

Send message
Joined: 14 Feb 04
Posts: 1817
Credit: 262,693,308
RAC: 44
Australia
Message 1179036 - Posted: 18 Dec 2011, 8:26:05 UTC - in response to Message 1179030.  

For those of you who've heard what I'm about to post many times over......<snip>

Not sure what your point was. At no time did I criticise the project in my OP I just made the point there was no use keeping the computers running when they were dry. It's a bit like leaving the car idling all day when the road's closed and you can't drive anywhere.

I don't necessarily think my connection problems are a SAH problem, although it "could" be due to the HE router still having brain seizures but I'm reasonably certain there is a problem somewhere in the PAIX (Palo Alto Internet(?) eXchange). However it needs pings and trace routes back in my direction to prove and locate it. I'm not the only one being effected by it either.

There, I've said it. Again. I feel better. Again.

Good, so do I :-)

T.A.
ID: 1179036 · Report as offensive
wsmith

Send message
Joined: 2 Nov 05
Posts: 4
Credit: 59,345,752
RAC: 0
United States
Message 1179058 - Posted: 18 Dec 2011, 12:27:26 UTC - in response to Message 1179030.  

I droped back to ver 6.10 from 6.12 and now I have all the units I can crunch.

i7 64bit
gtx580
win 7 64 bit
32 ram


when I went to 6.12 I got snake bit so I dropped back. My regular units are showing 6.03 and my cuda fermi is 6.10.


I am now crunching all day long and into the night

Dave
ID: 1179058 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 9786
Credit: 572,710,851
RAC: 3,799
Panama
Message 1179059 - Posted: 18 Dec 2011, 12:51:12 UTC - in response to Message 1179036.  

Hi

Something wierd happens here today, i switch the proxy from 69.163.96.2 (pacifica) to 184.22.244.105 (another california base) and all the slow DL and retrys simple disapears. Fast DL/UL and no error for more than an hour.

Simple coincidence or a prove the problem is in PAIX?.

Could anybody check it to?
ID: 1179059 · Report as offensive
Previous · 1 · 2 · 3

Message boards : Number crunching : Shutting Down the "Big Rigs"


 
©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.