recent woes


log in

Advanced search

Message boards : Technical News : recent woes

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · Next
Author Message
5subslr5
Send message
Joined: 4 Nov 02
Posts: 9
Credit: 11,434
RAC: 0
United States
Message 1039664 - Posted: 8 Oct 2010, 16:50:38 UTC - in response to Message 1038922.

"We know it will be fixed when it gets fixed." (Steve)

Deeeeply insightful !! heheheheheheeheh

Pascal Meeuws
Send message
Joined: 25 Nov 09
Posts: 5
Credit: 1,380,836
RAC: 0
Netherlands
Message 1039670 - Posted: 8 Oct 2010, 16:58:18 UTC

I've read the comments, and would like to give the people some advice.
Being a Software Engineer with database experience I can understand the amount of work the Seti people have to keep these systems up.

The best way to help them now is, throtlle back on the number of tasks being processed so that they can work on the issues at hand.

After the are solved, we can happily continue crunshing.

Regards,

Pascal

5subslr5
Send message
Joined: 4 Nov 02
Posts: 9
Credit: 11,434
RAC: 0
United States
Message 1039671 - Posted: 8 Oct 2010, 16:58:32 UTC - in response to Message 1039512.


.

There may be merit in what some have suggested in terms of shutting it all down until ways can be found to make Seti run reliably.




Merit in shut down and repair.......yes. Large merit....Yes. Massive merit....YES !

Profile Gary CharpentierProject donor
Volunteer tester
Avatar
Send message
Joined: 25 Dec 00
Posts: 12324
Credit: 6,624,443
RAC: 9,088
United States
Message 1039674 - Posted: 8 Oct 2010, 17:01:16 UTC - in response to Message 1039671.


.

There may be merit in what some have suggested in terms of shutting it all down until ways can be found to make Seti run reliably.




Merit in shut down and repair.......yes. Large merit....Yes. Massive merit....YES !

That could be several dozen years given the nature of funding ...
Still sure that is the way you want to go?

____________

5subslr5
Send message
Joined: 4 Nov 02
Posts: 9
Credit: 11,434
RAC: 0
United States
Message 1039687 - Posted: 8 Oct 2010, 17:21:22 UTC - in response to Message 1039674.
Last modified: 8 Oct 2010, 17:27:47 UTC


.

There may be merit in what some have suggested in terms of shutting it all down until ways can be found to make Seti run reliably.




Merit in shut down and repair.......yes. Large merit....Yes. Massive merit....YES !

That could be several dozen years given the nature of funding ...
Still sure that is the way you want to go?



Yup.

When continually trying the same thing has not worked there comes a point where something different may be in order.

msattlerProject donor
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38682
Credit: 573,496,580
RAC: 545,782
United States
Message 1039697 - Posted: 8 Oct 2010, 17:34:34 UTC

Server fund drive in progress......

Some $3000.00 so far.
The project will accept donations as little as $5.00.

Limber up the credit cards and go, boyz and gurlz.

Donate here.....now. Pleeeeeeeze?
____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

Arjant2
Send message
Joined: 14 Aug 10
Posts: 3
Credit: 52,475
RAC: 0
Netherlands
Message 1039701 - Posted: 8 Oct 2010, 17:37:10 UTC

I just switched from primegrid to Seti, to find out I cannot get any WU's.
Without having to read the entire page, is there anyone who can tell me how long it will take before the server is up again?

BarryAZ
Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 12,011,199
RAC: 4,649
United States
Message 1039705 - Posted: 8 Oct 2010, 17:42:04 UTC - in response to Message 1039687.

Right, which is why an alternate 'new approach' might be considered -- such as reducing the loads on the servers by redirecting a portion of the excessive workunit I/O into support of other worthy projects.

The 'keep doing the same thing even as it fails' routine applies as well to the 'SETI is my only project and when it doesn't work the way I want it, SETI is my only project' approach to problem solving.

Folks, figure this out, if you keep inundating this project with new work and completed work traffic, it isn't going to get to 'the good old days' (and frankly, in the life of the BOINC project history, SETI has never been the most reliable of projects, simply, by far, the largest). With the vastly increased CPU power and support for GPU processing, this project is up and over the wall and as long as folks go SETI only with their hardware, it is NOT going to change. I see that some group has developed an ATI GPU client for use with SETI, the good news is that it provides another source of computing power, the bad news is that it increases the overload potential for SETI.






When continually trying the same thing has not worked there comes a point where something different may be in order.


____________

msattlerProject donor
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38682
Credit: 573,496,580
RAC: 545,782
United States
Message 1039707 - Posted: 8 Oct 2010, 17:50:41 UTC - in response to Message 1039705.

There is nothing basically wrong with this project or it's approach that a little hardware cannot fix.

The bandwidth can be lived with, if we can just get the severs stable enough to keep it saturated. It always settles down once caches are filled.

Please consider donating to the cause. I have a donation drive going in the NC forum right now.

Even a $5.00 donation will be gratefully accepted.

You could do it in less time than it takes to post a dissenting opinion.

Please consider it.
____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

BarryAZ
Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 12,011,199
RAC: 4,649
United States
Message 1039758 - Posted: 8 Oct 2010, 18:42:44 UTC - in response to Message 1039707.
Last modified: 8 Oct 2010, 18:44:21 UTC

I think donating to this project is NOT a bad idea. I belive new hardware could certainly help somewhat. But I wonder -- perhaps the 'standard' half week outage in 'good times' suggests that not only is there a hardware need, but also, the amount of completed work is outstripping the capacity to process completed work in anything like real time - even with LOTS of new hardware.

My sense is that with new hardware (and I was under the impression that new hardware was 'in the pipeline'), the half-week off for processing will still be required. To the extent that is the case (and again, I am not absolutely certain this IS the case), certainly the Friday (upload and report) and Monday (fill the caches to overflowing) I/O jams will still exist (and frustrate folks who are SETI only BOINC players with a significant hardware investment in their own equipment).

So, I'd advocate a more comprehensive solution. Yes, toss money and hardware at the problem, but also, on the end user side, seriously consider sharing out the processing wealth. Certainly that share the processing wealth will make a lot of sense for the short run (as in weeks/months) until that day when the project has all the hardware that the project folks feel they need and can use.




There is nothing basically wrong with this project or it's approach that a little hardware cannot fix.

The bandwidth can be lived with, if we can just get the severs stable enough to keep it saturated. It always settles down once caches are filled.

Please consider donating to the cause. I have a donation drive going in the NC forum right now.

Even a $5.00 donation will be gratefully accepted.

You could do it in less time than it takes to post a dissenting opinion.

Please consider it.

____________

BarryAZ
Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 12,011,199
RAC: 4,649
United States
Message 1039763 - Posted: 8 Oct 2010, 18:49:08 UTC - in response to Message 1039701.

At a guess, and this is only a guess, assuming (and this is an assumption as the past few weeks have been quite problematic) thatthe I/O recovery jam up clears, project folks might be able to turn on the download servers later today or tomorrow.

But understand, because of the problems this project has been coping with over the past few weeks, a LOT of folks are going to be trying to get new work over the next several days.

You might also consider some other projects which support your CUDA GPU (what CUDA card do you have?). These include (depending on your GPU), Einstein, Collatz, Dnetc, and MilkyWay.


I just switched from primegrid to Seti, to find out I cannot get any WU's.
Without having to read the entire page, is there anyone who can tell me how long it will take before the server is up again?


____________

Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 15,204,196
RAC: 11,889
United States
Message 1039768 - Posted: 8 Oct 2010, 18:54:33 UTC - in response to Message 1039758.

(and I was under the impression that new hardware was 'in the pipeline'),

Oscar is on it's way but now another server is giving them fits. They were trying to contact some server manufacturers to see if they could talk them into a deep enough discount to use some of the money they got for Oscar to buy a second new server. This new fund drive is to make sure they can get what they need instead of just something to make do.
____________


PROUD MEMBER OF Team Starfire World BOINC

Arjant2
Send message
Joined: 14 Aug 10
Posts: 3
Credit: 52,475
RAC: 0
Netherlands
Message 1039801 - Posted: 8 Oct 2010, 19:32:27 UTC

Since I like the project and its goal, I have made a small donation.
I hope more people follow so they can get some new hardware and get things up and running again.

Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 15,204,196
RAC: 11,889
United States
Message 1039803 - Posted: 8 Oct 2010, 19:39:38 UTC - in response to Message 1039801.

Since I like the project and its goal, I have made a small donation.
I hope more people follow so they can get some new hardware and get things up and running again.


Come on over to the emergency fund thread http://setiathome.berkeley.edu/forum_thread.php?id=61691 and post in there. We are keeping a running total and thanking everyone for their donations. I believe the last time I peeked in there we were over $5000.

____________


PROUD MEMBER OF Team Starfire World BOINC

Profile ScarabDrowner
Volunteer tester
Avatar
Send message
Joined: 13 Sep 03
Posts: 90
Credit: 456,378
RAC: 0
United States
Message 1039898 - Posted: 8 Oct 2010, 22:39:00 UTC - in response to Message 1039803.

8.3 MILLION results returned and awaiting validation... ouch. No wonder things are still on standby.
____________

PhonAcq
Send message
Joined: 14 Apr 01
Posts: 1622
Credit: 22,065,802
RAC: 4,128
United States
Message 1039968 - Posted: 9 Oct 2010, 0:00:29 UTC - in response to Message 1039670.

I've read the comments, and would like to give the people some advice.
Being a Software Engineer with database experience I can understand the amount of work the Seti people have to keep these systems up.

The best way to help them now is, throtlle back on the number of tasks being processed so that they can work on the issues at hand.

After the are solved, we can happily continue crunshing.

Regards,

Pascal


I can't count the number of times I and others have made the same suggestion, to no avail. Berkeley in effect is limiting the number of tasks being issued by having a lot more downtime. It's an elegant solution in that no effort is required, freeing manpower to work on other things between crashes, but it is pretty lousy PR with the power crunchers.

PhonAcq
Send message
Joined: 14 Apr 01
Posts: 1622
Credit: 22,065,802
RAC: 4,128
United States
Message 1039975 - Posted: 9 Oct 2010, 0:11:29 UTC

It seems to be consensus that buying new hardware will fix all things. But I suspect that when the new hardware gets here, a new wave of issues will develop, which in turn will be blamed on not having state of the art hardware. And so on.

I really think more than a small effort should be in thinking about a better software solution to prevent these crashes. It is likely that scaling to more users, or more likely more powerful hosts (GPU's), has thrown a wrench in things and the system architecture needs to be revisited to handle an order of magnitude or more processing capacity than was originally envisioned. Would the arrival of the new server(s) be the time to start? Or is it that the software being used is not up to the tasks, and instead of hardware a better database needs to be purchased (or some such). It's worth debating at least.

[aside] This project is reminescent of MS-DOS and the early Windows versions. They kept adding functionality on an old dead horse, and got a way with it with good marketers and loyal apologists. (And I'm a Windows fan!)[/aside]

cwburch
Send message
Joined: 23 Jul 00
Posts: 51
Credit: 17,960,308
RAC: 20,057
United States
Message 1040043 - Posted: 9 Oct 2010, 2:16:43 UTC - in response to Message 1039154.

"I'm so old, that 98% of all trees in the forest, are younger than I am"

That's nothin'. I can remember when the dead sea got sick.
-Carl

Monty Chandler
Send message
Joined: 3 Apr 99
Posts: 6
Credit: 2,828,475
RAC: 623
United States
Message 1040055 - Posted: 9 Oct 2010, 2:29:41 UTC

Is the BOINC project still active? For a week or so everything stayed on my system. now it's all transferred but no new work is downloading. It seems the project is down more than up these days. I know the "Tech news" says it's scheduled to be down half the time - but that can't be right. No infrastructure is ever scheduled to be down 50% of the time for "maintenance". What's really going on? Change of staff?

I don't remember it being this bad in the past 11 years I've been churning results.. Perhaps it's just me getting old..


____________

Monty Chandler
Send message
Joined: 3 Apr 99
Posts: 6
Credit: 2,828,475
RAC: 623
United States
Message 1040059 - Posted: 9 Oct 2010, 2:36:35 UTC - in response to Message 1040055.

Never mind - I just read the thread...
____________

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · Next

Message boards : Technical News : recent woes

Copyright © 2014 University of California