## Panic Mode On (80) Server Problems?

Message boards : Number crunching : Panic Mode On (80) Server Problems?
Message board moderation

Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · 10 . . . 25 · Next

AuthorMessage
Mark Lybeck

Joined: 9 Aug 99
Posts: 244
Credit: 193,004,888
RAC: 41,843
Message 1323763 - Posted: 2 Jan 2013, 20:56:04 UTC - in response to Message 1323749.

I have over 150 machines that have not got any tasks today and they are getting tasks from Einstein which has a resource of 0.

Any idea why polling less than 303 seconds does not return the normal error about the request being too recent?

Um.. How do you fit all 150 machines at home? Or did you hijack a corporate server farm? How do you finance the powerbill for those 150 machines?

With a moderate assumption of 150W/machine would mean 22,5kW. At a 0,1€/kWh this would cost you 2,25€/hour -> 2,25€*24 = 54€/day -> 365* 54€/day = 19710€/year to run.

With the power youre spending you could trade in your car for a new one every year.

ID: 1323763 ·
HAL9000
Volunteer tester

Joined: 11 Sep 99
Posts: 6530
Credit: 185,777,170
RAC: 43,941
Message 1323765 - Posted: 2 Jan 2013, 21:02:00 UTC - in response to Message 1323763.

I have over 150 machines that have not got any tasks today and they are getting tasks from Einstein which has a resource of 0.

Any idea why polling less than 303 seconds does not return the normal error about the request being too recent?

Um.. How do you fit all 150 machines at home? Or did you hijack a corporate server farm? How do you finance the powerbill for those 150 machines?

With a moderate assumption of 150W/machine would mean 22,5kW. At a 0,1€/kWh this would cost you 2,25€/hour -> 2,25€*24 = 54€/day -> 365* 54€/day = 19710€/year to run.

With the power youre spending you could trade in your car for a new one every year.

IIRC he owns a server farm. Which he runs projects on when he doesn't have it leased out.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the BP6/VP6 User Group today!
ID: 1323765 ·
ivan
Volunteer tester

Joined: 5 Mar 01
Posts: 782
Credit: 290,568,370
RAC: 100,873
Message 1323768 - Posted: 2 Jan 2013, 21:07:19 UTC - in response to Message 1323757.

Guess I will just have to find some alcohol to get some patience, :^}

Yep, I know that feelin'... :-0!
ID: 1323768 ·
mikeej42

Joined: 26 Oct 00
Posts: 109
Credit: 791,863,759
RAC: 0
Message 1323773 - Posted: 2 Jan 2013, 21:22:28 UTC - in response to Message 1323763.

LOL
Yes these are servers at my work that I can use when we are not doing their "real work". I use them in between test cycles. They are HP c7000 Blade Chassis mounted 4 to a rack and the company was actually experimenting with special cooling systems for the equipment racks that the company was going to sell. There are 3 160KW cooling systems in our lab. The cooling system alone could handle several large homes. The company has since abandoned the cooling project.

ID: 1323773 ·
ivan
Volunteer tester

Joined: 5 Mar 01
Posts: 782
Credit: 290,568,370
RAC: 100,873
Message 1323789 - Posted: 2 Jan 2013, 21:38:23 UTC - in response to Message 1323773.

LOL
Yes these are servers at my work that I can use when we are not doing their "real work". I use them in between test cycles. They are HP c7000 Blade Chassis mounted 4 to a rack and the company was actually experimenting with special cooling systems for the equipment racks that the company was going to sell. There are 3 160KW cooling systems in our lab. The cooling system alone could handle several large homes. The company has since abandoned the cooling project.

I'm working on a project that's tooling up to use carbon-dioxide mixed-phase cooling in its system (particle detectors, not CPUs) but I'd love to commercialise it for processor cooling. Probably too expensive for home computing (a hefty compressor is needed to liquify the CO2) but maybe for server farms?
ID: 1323789 ·
Siran d'Vel'nahr
Volunteer tester

Joined: 23 May 99
Posts: 6223
Credit: 7,471,148
RAC: 76
Message 1323866 - Posted: 2 Jan 2013, 22:18:51 UTC

Greetings,

Dang near 5,000,000 WUs between MB and AstroPulse and my BOINC client event log says project has no new tasks?

What gives?

Keep on BOINCing...! :)

CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187

Windows 10 can be tamed into submission!

Microsoft wants us to "Imagine life without walls"...
I say, "If there are no walls, who needs Windows?"
ID: 1323866 ·
Qui-Gon
Volunteer tester

Joined: 15 May 99
Posts: 2932
Credit: 14,008,310
RAC: 6,870
Message 1323926 - Posted: 2 Jan 2013, 23:21:49 UTC - in response to Message 1323866.

Greetings,

Dang near 5,000,000 WUs between MB and AstroPulse and my BOINC client event log says project has no new tasks?

What gives?

Keep on BOINCing...! :)

Same thing happening to me. I took a break from crunching over the holidays so my hoppers are empty. Now that we're back up again, I can't get any work? That is disappointing.

PS: Happy New Year Siran
ID: 1323926 ·
Bernie Vine
Volunteer moderator
Volunteer tester

Joined: 26 May 99
Posts: 9289
Credit: 56,836,502
RAC: 16,268
Message 1323928 - Posted: 2 Jan 2013, 23:33:00 UTC

Just got 24 CPU tasks when I want GPU, when I looked all 24 are VLARS!!!

"Liberal and proud to be"
ID: 1323928 ·
mikeej42

Joined: 26 Oct 00
Posts: 109
Credit: 791,863,759
RAC: 0
Message 1323952 - Posted: 3 Jan 2013, 0:43:24 UTC - in response to Message 1323926.

Greetings,

Dang near 5,000,000 WUs between MB and AstroPulse and my BOINC client event log says project has no new tasks?

What gives?

Keep on BOINCing...! :)

Same thing happening to me. I took a break from crunching over the holidays so my hoppers are empty. Now that we're back up again, I can't get any work? That is disappointing.

PS: Happy New Year Siran

It appears that there are so many hosts with empty caches all trying to get tasks at the same time that the system is having troubles getting the tasks delivered. Tasks are coming out very slowly. I have gotten a few hundred between all my hosts today. We just need to remain patient while the servers get caught up.

Enjoy a mug of your favorite beverage and have a happy new year.
ID: 1323952 ·
.clair.

Joined: 4 Nov 04
Posts: 1300
Credit: 43,189,296
RAC: 20,883
Message 1323966 - Posted: 3 Jan 2013, 1:32:00 UTC

The cricket graph is well short of max,
somthing iz still broken.
the work cant get to the pipe to go slowly through it.
There shurly cant be that many VLARs not going to GPUs ??!!
I would realy like to get back to crunching VLARs on my ATI GPU but Eric an Co wont let me, [/whinge]
ID: 1323966 ·
zoom314
Volunteer tester

Joined: 30 Nov 03
Posts: 60535
Credit: 44,836,972
RAC: 5,813
Message 1323970 - Posted: 3 Jan 2013, 1:43:47 UTC - in response to Message 1323966.

The cricket graph is well short of max,
somthing iz still broken.
the work cant get to the pipe to go slowly through it.
There shurly cant be that many VLARs not going to GPUs ??!!
I would realy like to get back to crunching VLARs on my ATI GPU but Eric an Co wont let me, [/whinge]

Well I for one don't want a VLAR to impale My gpus...
What is BSG Robotech-Saga-Wiki SW-wiki
ID: 1323970 ·
Grant (SSSF)
Volunteer tester

Joined: 19 Aug 99
Posts: 9910
Credit: 128,731,791
RAC: 82,120
Message 1323991 - Posted: 3 Jan 2013, 3:47:43 UTC - in response to Message 1323966.

somthing iz still broken.

Yep.
If there are that many WUs ready to send, and everything was working as it should then the traffic should be maxed out. It's not, and most requests for work result in the No tasks available message.

EDIT- one of my systems just scored 38 WUs for the GPU. They're not shorties, so it should last for a few hours before they're all done.
Grant
Darwin NT
ID: 1323991 ·
ExchangeMan
Volunteer tester

Joined: 9 Jan 00
Posts: 115
Credit: 157,719,104
RAC: 0
Message 1323995 - Posted: 3 Jan 2013, 4:08:15 UTC - in response to Message 1323991.

I just got 44 GPU work units - unfortunately all shorties.

ID: 1323995 ·
Peter M. Ferrie
Volunteer tester

Joined: 28 Mar 03
Posts: 86
Credit: 9,951,230
RAC: 0
Message 1323998 - Posted: 3 Jan 2013, 4:27:58 UTC

if there really are 5,000,000 + workunits ready to send ... it will be the great wu flood of 2013 .. head for higher cpu's
ID: 1323998 ·
Keith White

Joined: 29 May 99
Posts: 390
Credit: 7,211,318
RAC: 8,793
Message 1324000 - Posted: 3 Jan 2013, 4:40:47 UTC

I've gotten 26 so far.

Back of the envelope calculations suggest at a rate of 25 units per second could be downloaded, limited due to the pipe's bandwidth. This guesstimate doesn't include any outgoing bandwidth used to for control and status requests.

Now since a significant number of us are using GPUs to help do the heavy lifting, a two day outage would have wiped out most if not all of our 100 unit queue so everybody wants 100 GPU units. That comes out to only filling 900 hosts per hour, best case. So how many active hosts out there crunching with one or more GPUs? 10,000? 20,000? So it's going to take a dozen hours or two to handle demand and that's doesn't include the fact that for the fastest among us could crunch 100 GPU units in under an hour and a good number of us in under a day. It'll take a bit.

And since I've started writing this, I got 41 more.
"Life is just nature's way of keeping meat fresh." - The Doctor
ID: 1324000 ·
Grant (SSSF)
Volunteer tester

Joined: 19 Aug 99
Posts: 9910
Credit: 128,731,791
RAC: 82,120
Message 1324007 - Posted: 3 Jan 2013, 5:18:24 UTC - in response to Message 1324000.

So it's going to take a dozen hours or two to handle demand and that's doesn't include the fact that for the fastest among us could crunch 100 GPU units in under an hour and a good number of us in under a day. It'll take a bit.

My GTX 560Ti does roughly 200 WUs per day- if they are all 22min crunching time. Some are longer, most are shorter.
On average it probably does closer to 300WU per day. When it can get them.

Grant
Darwin NT
ID: 1324007 ·
Wiggo "Socialist"

Joined: 24 Jan 00
Posts: 14263
Credit: 185,604,891
RAC: 82,280
Message 1324036 - Posted: 3 Jan 2013, 7:53:19 UTC - in response to Message 1324007.

It seems that the feeder is having a bit of a problem feeding us.

Cheers.
ID: 1324036 ·
Bernie Vine
Volunteer moderator
Volunteer tester

Joined: 26 May 99
Posts: 9289
Credit: 56,836,502
RAC: 16,268
Message 1324041 - Posted: 3 Jan 2013, 8:14:06 UTC

Still just getting VLARS for the CPU. Never seen this many at once. Nothing for GPU.
"Liberal and proud to be"
ID: 1324041 ·
Keith White

Joined: 29 May 99
Posts: 390
Credit: 7,211,318
RAC: 8,793
Message 1324051 - Posted: 3 Jan 2013, 9:35:52 UTC - in response to Message 1324041.

Well so far I've gotten all of 67 MB units. 56 GPU, 11 CPU, none VLAR.

I'm now at 99 GPU and 46 CPU.

Like I said before, best super ideal guesstimate is that maybe 900 hosts can be fed 100 units per hour. At the current cricket levels I'll guess maybe 500 hosts per hour.

So how many hosts want GPU units? How many do an Oliver Twist and ask for another full reload in 3/6/12/24 hours?
"Life is just nature's way of keeping meat fresh." - The Doctor
ID: 1324051 ·
kittyman
Volunteer tester

Joined: 9 Jul 00
Posts: 49862
Credit: 915,122,501
RAC: 160,616
Message 1324052 - Posted: 3 Jan 2013, 9:42:36 UTC - in response to Message 1324041.

Still just getting VLARS for the CPU. Never seen this many at once. Nothing for GPU.

I set 'use nVidia GPU' to no for the night, will try to build some CPU cache.
Otherwise the rigs were just asking for GPU and getting NOTHING.
I'll enable GPU in the morning and see if anything's changed.
What meowing lurks in the hearts of man? The kittyman knows....MEOWhahahahahahha!