Panic Mode On (82) Server Problems?

Message boards : Number crunching : Panic Mode On (82) Server Problems?
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 16 · 17 · 18 · 19 · 20 · 21 · 22 . . . 24 · Next

AuthorMessage
Thomas
Volunteer tester

Send message
Joined: 9 Dec 11
Posts: 1499
Credit: 1,345,576
RAC: 0
France
Message 1354034 - Posted: 6 Apr 2013, 8:53:01 UTC - in response to Message 1354032.  

RIP Compaq P1100... Sniff... :(
ID: 1354034 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1354036 - Posted: 6 Apr 2013, 8:55:34 UTC - in response to Message 1354032.  
Last modified: 6 Apr 2013, 8:57:51 UTC

I'm going to try opening this one up in the next day or two and see if it is just a capacitor or 10 that I can replace.

90% chance.
Ok when warm, but unhappy when cold is symptomatic of dying electros. With the symptoms you've described i'd shotgun the caps in the power supply & horizontal output & drive sections.

EDIT- use 105°c low ESR caps.
Grant
Darwin NT
ID: 1354036 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1354058 - Posted: 6 Apr 2013, 10:30:39 UTC - in response to Message 1354036.  

I'm going to try opening this one up in the next day or two and see if it is just a capacitor or 10 that I can replace.

90% chance.
Ok when warm, but unhappy when cold is symptomatic of dying electros. With the symptoms you've described i'd shotgun the caps in the power supply & horizontal output & drive sections.

EDIT- use 105°c low ESR caps.

Well.. ambient temperature has nothing to do with it. The electronics themselves need to be toasty for it to work fine. I've figured it was capacitors for a while, I was just hoping it would hold on for dear life for as long as possible.

When I was power cycling it, following the screech and the power-off click 2 seconds later, the epic amount of static crackling/popping from inside is amazing. Best description I have for that is running over a sheet of bubble wrap with a steamroller.

I'm going to see if I can fix it.. but the new monitor has been ordered. Cashed in some IOUs. So fixing will be a side-project with no rush.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1354058 · 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 1354061 - Posted: 6 Apr 2013, 10:37:17 UTC - in response to Message 1354036.  

I'm going to try opening this one up in the next day or two and see if it is just a capacitor or 10 that I can replace.

90% chance.
Ok when warm, but unhappy when cold is symptomatic of dying electros. With the symptoms you've described i'd shotgun the caps in the power supply & horizontal output & drive sections.

EDIT- use 105°c low ESR caps.

Wow....nice one, dude.

I did not think anybody else knew what a low ESR cap was.
I sold industrial/commercial parts for many years.
When guys were shotgunning their personal projects, I always told them...hey, buy the black ones, not the blue ones.

Luv 'dem 105's.

"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1354061 · Report as offensive
Profile shizaru
Volunteer tester
Avatar

Send message
Joined: 14 Jun 04
Posts: 1130
Credit: 1,967,904
RAC: 0
Greece
Message 1354081 - Posted: 6 Apr 2013, 13:25:50 UTC - in response to Message 1354032.  
Last modified: 6 Apr 2013, 13:28:15 UTC

I'm going to try opening this one up in the next day or two and see if it is just a capacitor or 10 that I can replace.


Good luck! Highly recommended course of action BTW:) Had great success with an LG monitor brought back to life by 2x50 (euro)cent caps. The six year plus timeframe fits too. Wasn't that around the time of The Great Cap Plague?

Try Googleing your model number + "cap replacement". Worked for me:) (Only one of the two was bulging)
ID: 1354081 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1354141 - Posted: 6 Apr 2013, 17:54:35 UTC - in response to Message 1354081.  
Last modified: 6 Apr 2013, 17:54:58 UTC

Wow....nice one, dude.

I did not think anybody else knew what a low ESR cap was.
I sold industrial/commercial parts for many years.
When guys were shotgunning their personal projects, I always told them...hey, buy the black ones, not the blue ones.

Luv 'dem 105's.


I've done this before, if you may remember.. Anyone ever have to replace capacitors? Though I don't recall if I used low ESR or not.. but the replacement caps were 125C.

I'm going to try opening this one up in the next day or two and see if it is just a capacitor or 10 that I can replace.


Good luck! Highly recommended course of action BTW:) Had great success with an LG monitor brought back to life by 2x50 (euro)cent caps. The six year plus timeframe fits too. Wasn't that around the time of The Great Cap Plague?

Try Googleing your model number + "cap replacement". Worked for me:) (Only one of the two was bulging)


I've had the monitor for six years, but it was manufactured in June 1998, so says the sticker on the back of it. I plan on seeing if I can revive it. I found the spec page for it, too. I used to use it on the lower resolutions at the 100+ refresh rate, but sometime around when the nVidia drivers got into the ~180.xx range, they dropped support for anything over 85Hz. I just know I used to play some of my games in 640x480 @ 160 Hz. That was a smooth picture, let me tell you...
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1354141 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1354178 - Posted: 6 Apr 2013, 21:08:16 UTC - in response to Message 1354023.  

Scheduler appears to be having issues again.
No problems for several hours, then for several requests for work (while reporting work done) it says i've reached the limits. After 3-10 requests it suddenly decides i haven't reached my limit & gives out a whole bunch in one hit.
...

It's possible that's intentional, a change in limits can be applied at any time. Perhaps Matt has a script which uses it to throttle things as needed.
                                                                   Joe
ID: 1354178 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1354196 - Posted: 6 Apr 2013, 21:52:01 UTC - in response to Message 1354178.  

Scheduler appears to be having issues again.
No problems for several hours, then for several requests for work (while reporting work done) it says i've reached the limits. After 3-10 requests it suddenly decides i haven't reached my limit & gives out a whole bunch in one hit.
...

It's possible that's intentional, a change in limits can be applied at any time. Perhaps Matt has a script which uses it to throttle things as needed.
                                                                   Joe

I don't think so- it seems to be occuring at random times- sometimes when there's a huge load, other times when things are almost dead (relatively speaking).
Grant
Darwin NT
ID: 1354196 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1354198 - Posted: 6 Apr 2013, 21:54:34 UTC - in response to Message 1354196.  


I notice the network traffic has dropped right off- no more AP available to split at the moment. After a slight dip down to 84Mb/s it's back up to 100Mb/s.
Grant
Darwin NT
ID: 1354198 · 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 1354204 - Posted: 6 Apr 2013, 22:10:08 UTC

Caches on all 9 rigs are full up here...
Could just be the lack of AP going out and possibly caches are finally satisfied on most hosts.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1354204 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34763
Credit: 261,360,520
RAC: 489
Australia
Message 1354205 - Posted: 6 Apr 2013, 22:18:01 UTC - in response to Message 1354204.  

Caches on all 9 rigs are full up here...
Could just be the lack of AP going out and possibly caches are finally satisfied on most hosts.

Those AP's being shoved out at high rates were a really big headache with the old link.

Cheers.
ID: 1354205 · 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 1354206 - Posted: 6 Apr 2013, 22:20:04 UTC - in response to Message 1354205.  

Caches on all 9 rigs are full up here...
Could just be the lack of AP going out and possibly caches are finally satisfied on most hosts.

Those AP's being shoved out at high rates were a really big headache with the old link.

Cheers.

Yes, they were. But not anymore.

What made it worse was when they went into backoff land, Boinc would not ask for any more new work until they made it through.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1354206 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1354207 - Posted: 6 Apr 2013, 22:37:55 UTC - in response to Message 1354206.  


Now that AP has stopped, and the shorties have gone through it looks like the baseline is around 90Mb/s. Given that 80% of your total bandwidth is the absolute maximum you want to use (80Mb/s on the old link) it's easy to see why the network was falling over itself even at the quietest times.
And looking at the current graphs when AP & the shorties were going through there were extended periods of 185-190Mb/s and many surges of over 200MB/s- with one peak of 312MB/s (for some reason there's long been a big sharp surge of traffic after 0:00HRs Berkeley time).

So as things stand 250MB/s appears to be the absolute minimum the project needs at the moment. Even with v7 coming out with longer crunch times- another generation or 2 of GPU, the price of the current ones dropping like a stone, and the return of crunchers that moved to other projects due to the problems of getting work returning now it's readily available i can see that 250MB/s minimum being not nearly enough within 12 months.
Grant
Darwin NT
ID: 1354207 · 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 1354212 - Posted: 6 Apr 2013, 22:57:48 UTC - in response to Message 1354207.  

Well....all I can tell ya is right now, the kitties are very happy campers.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1354212 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9954
Credit: 103,452,613
RAC: 328
United Kingdom
Message 1354215 - Posted: 6 Apr 2013, 23:09:58 UTC - in response to Message 1354207.  


So as things stand 250MB/s appears to be the absolute minimum the project needs at the moment. Even with v7 coming out with longer crunch times- another generation or 2 of GPU, the price of the current ones dropping like a stone, and the return of crunchers that moved to other projects due to the problems of getting work returning now it's readily available i can see that 250MB/s minimum being not nearly enough within 12 months.

As has been said before, Eric made this comment over on Beta

Our new network connection should be able to handle about 5 times the bandwidth of the old one. We're currently running it at about half capacity.

So hopefully we are good up to 500MB/s
ID: 1354215 · Report as offensive
Profile RottenMutt
Avatar

Send message
Joined: 15 Mar 01
Posts: 1011
Credit: 230,314,058
RAC: 0
United States
Message 1354216 - Posted: 6 Apr 2013, 23:13:33 UTC - in response to Message 1354198.  
Last modified: 6 Apr 2013, 23:17:08 UTC


I notice the network traffic has dropped right off- no more AP available to split at the moment. After a slight dip down to 84Mb/s it's back up to 100Mb/s.

i noticed to, tapes are done. usually i get a good mix of AP, but not since the co-lo.
ID: 1354216 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1354218 - Posted: 6 Apr 2013, 23:22:19 UTC - in response to Message 1354215.  


So as things stand 250MB/s appears to be the absolute minimum the project needs at the moment. Even with v7 coming out with longer crunch times- another generation or 2 of GPU, the price of the current ones dropping like a stone, and the return of crunchers that moved to other projects due to the problems of getting work returning now it's readily available i can see that 250MB/s minimum being not nearly enough within 12 months.

As has been said before, Eric made this comment over on Beta

Our new network connection should be able to handle about 5 times the bandwidth of the old one. We're currently running it at about half capacity.

So hopefully we are good up to 500MB/s

Yep.
So when they use the network for moving archive data around it should only take a few hours, not days, and not even have any effect on uploads & downloads.


And i just realised it's going to take a while to get used to the new graphs- i just had a look & thought things were broken again as it's dropped so low- low being 95Mb/s....
Grant
Darwin NT
ID: 1354218 · Report as offensive
Profile Gary Charpentier Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 25 Dec 00
Posts: 30651
Credit: 53,134,872
RAC: 32
United States
Message 1354239 - Posted: 7 Apr 2013, 1:57:19 UTC - in response to Message 1354215.  


So as things stand 250MB/s appears to be the absolute minimum the project needs at the moment. Even with v7 coming out with longer crunch times- another generation or 2 of GPU, the price of the current ones dropping like a stone, and the return of crunchers that moved to other projects due to the problems of getting work returning now it's readily available i can see that 250MB/s minimum being not nearly enough within 12 months.

As has been said before, Eric made this comment over on Beta

Our new network connection should be able to handle about 5 times the bandwidth of the old one. We're currently running it at about half capacity.

So hopefully we are good up to 500MB/s

IIRC some time ago Matt commented they did a test run and the servers could not do more than just under 500MB/s due to disk I/O becoming saturated. I suspect that is the limit Eric is talking about and not the physical link speed.

I'm sure they will slowly be raising the imposed limits over time, making sure that no issues come up.

ID: 1354239 · 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 1354502 - Posted: 7 Apr 2013, 22:33:03 UTC

Well, one new dataset was loaded and is now splitting AP.
The Crickets responded by jumping up to around 140Mb/s.

Not hard to see why the servers got so tangled up so fast when AP hit the pipe with a 100Mb cap.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1354502 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1354504 - Posted: 7 Apr 2013, 22:46:52 UTC - in response to Message 1354502.  
Last modified: 7 Apr 2013, 22:47:02 UTC

Well, one new dataset was loaded and is now splitting AP.
The Crickets responded by jumping up to around 140Mb/s.

Not hard to see why the servers got so tangled up so fast when AP hit the pipe with a 100Mb cap.

I want to see a Shortie Storm for comparison.

Claggy
ID: 1354504 · Report as offensive
Previous · 1 . . . 16 · 17 · 18 · 19 · 20 · 21 · 22 . . . 24 · Next

Message boards : Number crunching : Panic Mode On (82) 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.