Panic Mode On (28) Server problems

Message boards : Number crunching : Panic Mode On (28) Server problems
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 16 · Next

AuthorMessage
Rasputin
Volunteer tester

Send message
Joined: 13 Jun 02
Posts: 1764
Credit: 6,132,221
RAC: 0
Russia
Message 966872 - Posted: 30 Jan 2010, 8:56:07 UTC - in response to Message 966870.  

Present batch of Shorties is sure maxing out the pipe.

Dave


I just dumped Vista and went back to windows XP. Updated everything, installed BOINC...so far so good.. Then CUDA units were completing in 4-5 minutes! Thought for sure something was messed up but apparently not if your getting shorties too.

I was about to reinstall BOINC.. Anyway, thanks for mentioning that. Saved me a lot of time and frustration.
ID: 966872 · Report as offensive
_heinz
Volunteer tester

Send message
Joined: 25 Feb 05
Posts: 744
Credit: 5,539,270
RAC: 0
France
Message 966903 - Posted: 30 Jan 2010, 12:34:11 UTC

30.01.2010 13:27:50 SETI@home Temporarily failed download of ap_01ja07ae_B5_P0_00193_20100129_23450.wu: HTTP error
30.01.2010 13:27:50 SETI@home Backing off 1 hr 21 min 8 sec on download of ap_01ja07ae_B5_P0_00193_20100129_23450.wu
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
get two astropulse, but second hung up....
any ideas
ID: 966903 · Report as offensive
Profile Keith T.
Volunteer tester
Avatar

Send message
Joined: 23 Aug 99
Posts: 962
Credit: 537,293
RAC: 9
United Kingdom
Message 966906 - Posted: 30 Jan 2010, 12:51:13 UTC - in response to Message 966903.  
Last modified: 30 Jan 2010, 12:56:24 UTC

http://fragment1.berkeley.edu/newcricket/grapher.cgi?target=%2Frouter-interfaces%2Finr-250%2Fgigabitethernet2_3;ranges=d%3Aw%3Am%3Ay;view=octets has been pegged near the top for the last 18 hours or so, and has been higher than recent average for the last few days.

This is probably due to the combination of the return of Astropulse, combined with another run of "shorties". Best advice is probably to just be patient.

[edit]
Those graphs are Berkeley local time (PST), not UTC.
ID: 966906 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51478
Credit: 1,018,363,574
RAC: 1,004
United States
Message 966910 - Posted: 30 Jan 2010, 13:03:29 UTC - in response to Message 966906.  

http://fragment1.berkeley.edu/newcricket/grapher.cgi?target=%2Frouter-interfaces%2Finr-250%2Fgigabitethernet2_3;ranges=d%3Aw%3Am%3Ay;view=octets has been pegged near the top for the last 18 hours or so, and has been higher than recent average for the last few days.

This is probably due to the combination of the return of Astropulse, combined with another run of "shorties". Best advice is probably to just be patient.

[edit]
Those graphs are Berkeley local time (PST), not UTC.


Yup.
The crickets are chirping at warp speed again........

"Time is simply the mechanism that keeps everything from happening all at once."

ID: 966910 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 966911 - Posted: 30 Jan 2010, 13:07:25 UTC - in response to Message 966906.  

This is probably due to the combination of the return of Astropulse, combined with another run of "shorties". Best advice is probably to just be patient.

Agreed on all counts. I got an allocation of 50 shorties in one request a while back, so I've been down in the crunch farm watching how the improved backoff/retry mechanism in the newer clients handles the clogged pipe (first real test for a long time). Looks pretty good to me.
ID: 966911 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51478
Credit: 1,018,363,574
RAC: 1,004
United States
Message 966914 - Posted: 30 Jan 2010, 13:14:13 UTC - in response to Message 966911.  
Last modified: 30 Jan 2010, 13:19:11 UTC

This is probably due to the combination of the return of Astropulse, combined with another run of "shorties". Best advice is probably to just be patient.

Agreed on all counts. I got an allocation of 50 shorties in one request a while back, so I've been down in the crunch farm watching how the improved backoff/retry mechanism in the newer clients handles the clogged pipe (first real test for a long time). Looks pretty good to me.

I got a real good cache on all 3 Cuda rigs, so I am like 10 days or more behind the curve on this one.....

I absolutely hate the way the newer clients handle a clogged pipe. The 'project backoffs' are a PITA.......

It's hard to even push the buttons enough to get them to push up and downloads through when the servers got their undies in a bundle.

May be good for the servers, but it's hell for the clients.

At least the old versions will still chink away at the pending transactions and push them through one at a time as they can connect.

EDIT.....
And no, plinking the servers constantly is NOT my problem....
It is the SERVER'S problem.

Now we are up to the old 'bandwidth up the hill' argument.
I personally don't give one rat's ass about it, as long as the pipe stays tight at 100mbs or whatever it can do, I have as good a chance as anybody at connecting.
An upgrade would be nice, but would the servers be able to handle it then?

It's all a pig in a poke unless some real hard choices are made up the hill. And the money to implement them.
"Time is simply the mechanism that keeps everything from happening all at once."

ID: 966914 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 966979 - Posted: 30 Jan 2010, 17:29:03 UTC - in response to Message 966914.  


I absolutely hate the way the newer clients handle a clogged pipe. The 'project backoffs' are a PITA.......

It's hard to even push the buttons enough to get them to push up and downloads through when the servers got their undies in a bundle.

May be good for the servers, but it's hell for the clients.

At least the old versions will still chink away at the pending transactions and push them through one at a time as they can connect.

EDIT.....
And no, plinking the servers constantly is NOT my problem....
It is the SERVER'S problem.

Now we are up to the old 'bandwidth up the hill' argument.
I personally don't give one rat's ass about it, as long as the pipe stays tight at 100mbs or whatever it can do, I have as good a chance as anybody at connecting.
An upgrade would be nice, but would the servers be able to handle it then?

It's all a pig in a poke unless some real hard choices are made up the hill. And the money to implement them.

So, let me make sure I understand this.

If your sewer line is clogged, you solve the problem by walking around the house, turning on all the faucets, and flushing all the toilets.

Maybe inviting over friends and getting a keg of beer.

... and it isn't your problem if everything overflows.

Yet, if you were careful about using the drains, you'd save the floor, and could call the plumber during normal business hours and not pay extra.

Uploading and downloading? Same thing.
ID: 966979 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 966991 - Posted: 30 Jan 2010, 17:59:10 UTC - in response to Message 966979.  


I absolutely hate the way the newer clients handle a clogged pipe. The 'project backoffs' are a PITA.......

It's hard to even push the buttons enough to get them to push up and downloads through when the servers got their undies in a bundle.

May be good for the servers, but it's hell for the clients.

At least the old versions will still chink away at the pending transactions and push them through one at a time as they can connect.

EDIT.....
And no, plinking the servers constantly is NOT my problem....
It is the SERVER'S problem.

Now we are up to the old 'bandwidth up the hill' argument.
I personally don't give one rat's ass about it, as long as the pipe stays tight at 100mbs or whatever it can do, I have as good a chance as anybody at connecting.
An upgrade would be nice, but would the servers be able to handle it then?

It's all a pig in a poke unless some real hard choices are made up the hill. And the money to implement them.

So, let me make sure I understand this.

If your sewer line is clogged, you solve the problem by walking around the house, turning on all the faucets, and flushing all the toilets.

Maybe inviting over friends and getting a keg of beer.

... and it isn't your problem if everything overflows.

Yet, if you were careful about using the drains, you'd save the floor, and could call the plumber during normal business hours and not pay extra.

Uploading and downloading? Same thing.

Sounds like a breaking the Lease Party I had many many years ago.....LOL
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 966991 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51478
Credit: 1,018,363,574
RAC: 1,004
United States
Message 966997 - Posted: 30 Jan 2010, 18:10:28 UTC - in response to Message 966979.  



Uploading and downloading? Same thing.


No......I just keep pissing up the rope and hope it does not trickle back down......
And hope whoever is holding the rope up top pulls it in.

"Time is simply the mechanism that keeps everything from happening all at once."

ID: 966997 · Report as offensive
Dave

Send message
Joined: 29 Mar 02
Posts: 778
Credit: 25,001,396
RAC: 0
United Kingdom
Message 967014 - Posted: 30 Jan 2010, 19:06:32 UTC

I must admit to kicking things manually every time 1 of my offline clients wants a dump (so to speak).
ID: 967014 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 967016 - Posted: 30 Jan 2010, 19:17:15 UTC - in response to Message 967014.  

I must admit to kicking things manually every time 1 of my offline clients wants a dump (so to speak).

Boy I have Never done that...(Watching out for Ned)LOL
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 967016 · Report as offensive
FiveHamlet
Avatar

Send message
Joined: 5 Oct 99
Posts: 783
Credit: 32,638,578
RAC: 0
United Kingdom
Message 967017 - Posted: 30 Jan 2010, 19:17:39 UTC - in response to Message 966872.  

Your Welcome.

Dave
ID: 967017 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 967062 - Posted: 30 Jan 2010, 21:19:58 UTC - in response to Message 966997.  



Uploading and downloading? Same thing.


No......I just keep pissing up the rope and hope it does not trickle back down......
And hope whoever is holding the rope up top pulls it in.

Interesting that you left out everything but the least interesting part.

Either analogy sounds like you're going to make a mess, and then have to deal with it.
ID: 967062 · Report as offensive
Rasputin
Volunteer tester

Send message
Joined: 13 Jun 02
Posts: 1764
Credit: 6,132,221
RAC: 0
Russia
Message 967067 - Posted: 30 Jan 2010, 21:23:49 UTC - in response to Message 967017.  

Your Welcome.

Dave


I guess the shorties have run dry. All WU's are crunching in normal times now.
ID: 967067 · Report as offensive
FiveHamlet
Avatar

Send message
Joined: 5 Oct 99
Posts: 783
Credit: 32,638,578
RAC: 0
United Kingdom
Message 967075 - Posted: 30 Jan 2010, 21:29:23 UTC - in response to Message 967067.  

Rig on a Bench still has 870 to do at the moment.
At 1 every 45 secs that's another 11 hrs or so. LOL.

Dave
ID: 967075 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 967079 - Posted: 30 Jan 2010, 21:31:40 UTC - in response to Message 967075.  

Rig on a Bench still has 870 to do at the moment.
At 1 every 45 secs that's another 11 hrs or so. LOL.

Dave

What are those puppies running at? Mine seem to be around 2 minutes each.
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 967079 · Report as offensive
FiveHamlet
Avatar

Send message
Joined: 5 Oct 99
Posts: 783
Credit: 32,638,578
RAC: 0
United Kingdom
Message 967086 - Posted: 30 Jan 2010, 21:39:59 UTC - in response to Message 967079.  
Last modified: 30 Jan 2010, 21:41:34 UTC

Sorry for misinformation 4 GPU's so in 3 mins thats 1 every 45secs
Everything is stock here.

Dave
ID: 967086 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 967092 - Posted: 30 Jan 2010, 22:01:09 UTC - in response to Message 966914.  


And no, plinking the servers constantly is NOT my problem....
It is the SERVER'S problem.

Now we are up to the old 'bandwidth up the hill' argument.
I personally don't give one rat's ass about it, as long as the pipe stays tight at 100mbs or whatever it can do, I have as good a chance as anybody at connecting.
An upgrade would be nice, but would the servers be able to handle it then?

It's all a pig in a poke unless some real hard choices are made up the hill. And the money to implement them.

The most cost-effective solution to the "plinking the servers" problem is to make a BOINC client that doesn't plink.

Same with your 100 megabit question (peak throughput is probably about 90 megabits). If the clients could tell that the wire was at 90 megabits and take turns more slowly then *throughput* would go up.

But BOINC features that do that (like the project-wide backoff) are widely condemned even though they do exactly what you say you want.
ID: 967092 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51478
Credit: 1,018,363,574
RAC: 1,004
United States
Message 967134 - Posted: 31 Jan 2010, 1:01:08 UTC - in response to Message 967092.  


And no, plinking the servers constantly is NOT my problem....
It is the SERVER'S problem.

Now we are up to the old 'bandwidth up the hill' argument.
I personally don't give one rat's ass about it, as long as the pipe stays tight at 100mbs or whatever it can do, I have as good a chance as anybody at connecting.
An upgrade would be nice, but would the servers be able to handle it then?

It's all a pig in a poke unless some real hard choices are made up the hill. And the money to implement them.

The most cost-effective solution to the "plinking the servers" problem is to make a BOINC client that doesn't plink.

Same with your 100 megabit question (peak throughput is probably about 90 megabits). If the clients could tell that the wire was at 90 megabits and take turns more slowly then *throughput* would go up.

But BOINC features that do that (like the project-wide backoff) are widely condemned even though they do exactly what you say you want.


Maybe I did not state my case clearly.......

I do NOT want the client to do a project wide backoff when it feels it necessary.

I want it the plonk the servers anytime it needs to get work or report it.

I know this is hard on the servers and bandwidth....

But that is what I want. When the servers are in trouble and it takes a couple of DAYS to clear the ready to send buffer, it drives the kitties wild.

The older client will keep trying, one at a time, until things get through.
The newer client will just back off like a lazy stepchild, and quit even trying just because a few attempts don't make it.

"Time is simply the mechanism that keeps everything from happening all at once."

ID: 967134 · Report as offensive
Profile 52 Aces
Avatar

Send message
Joined: 7 Jan 02
Posts: 497
Credit: 14,261,068
RAC: 67
United States
Message 967136 - Posted: 31 Jan 2010, 1:10:47 UTC - in response to Message 967134.  

I know this is hard on the servers and bandwidth....

But that is what I want.


That's what the [Retry Now] and [Update] buttons are for. Oh, too human labor intensive? Fair is fair... ;-)
ID: 967136 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 16 · Next

Message boards : Number crunching : Panic Mode On (28) Server problems


 
©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.