Out of the fire and into the pit of sulfuric acid. (Feb 19, 2010)


log in

Advanced search

Message boards : Technical News : Out of the fire and into the pit of sulfuric acid. (Feb 19, 2010)

Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 · Next
Author Message
Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 15,507,401
RAC: 11,344
United States
Message 973302 - Posted: 22 Feb 2010, 21:46:16 UTC - in response to Message 973298.

Yes, I say again, WELL DONE!!! Bet we are really hammering the servers, hope we don't break anything else! :-)
____________


PROUD MEMBER OF Team Starfire World BOINC

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8491
Credit: 49,780,825
RAC: 53,944
United Kingdom
Message 973304 - Posted: 22 Feb 2010, 21:50:25 UTC

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
b) Tomorrow, while things are quiet during mainteance

Be good to get some 'normal case' reports over the same paths as the existing tests, to predict whether the technique will be any use during the next outage.

Profile Siran d'Vel'nahr
Volunteer tester
Avatar
Send message
Joined: 23 May 99
Posts: 5689
Credit: 4,697,365
RAC: 2,790
United States
Message 973305 - Posted: 22 Feb 2010, 21:53:16 UTC - in response to Message 973302.

Yes, I say again, WELL DONE!!! Bet we are really hammering the servers, hope we don't break anything else! :-)

Yes, I agree. I still have 33 WUs, to upload, on my Windoze box. I'm hoping they can get here and get reported before all the "hammering" does indeed break something. ;)

Keep on BOINCing...! :)

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

Siran's website: [ ONLINE! ]

Ranger
Send message
Joined: 27 Dec 00
Posts: 49
Credit: 39,318,995
RAC: 7,393
United States
Message 973308 - Posted: 22 Feb 2010, 22:04:23 UTC - in response to Message 973304.

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
b) Tomorrow, while things are quiet during mainteance

Be good to get some 'normal case' reports over the same paths as the existing tests, to predict whether the technique will be any use during the next outage.


I wrote and scheduled a script that has been tracking pathping stats every hour since 2010-02-20 16:00 UTC. I may let it run through 2010-03-09 to get a feel for "normal".

____________

Profile Dan Rhodes
Avatar
Send message
Joined: 28 Jul 04
Posts: 56
Credit: 865,942
RAC: 0
United Kingdom
Message 973315 - Posted: 22 Feb 2010, 22:14:40 UTC
Last modified: 22 Feb 2010, 22:15:07 UTC

Poor Upload servers, they are getting a right hammering!

http://fragment1.berkeley.edu/newcricket/grapher.cgi?target=%2Frouter-interfaces%2Finr-250%2Fgigabitethernet2_3;view=Octets;ranges=d

Fingers crossed no disks go!

D
____________
Cheers
D

Mikek69
Send message
Joined: 18 Sep 04
Posts: 3
Credit: 175,782
RAC: 0
United Kingdom
Message 973325 - Posted: 22 Feb 2010, 22:26:28 UTC

Interesting. WHile you lot were frozen out I was getting downloads on cue as WUs finished. Just the upload and reporting were a problem. Not now though, all gone :D

archae86
Send message
Joined: 31 Aug 99
Posts: 888
Credit: 1,572,688
RAC: 5
United States
Message 973351 - Posted: 22 Feb 2010, 23:59:08 UTC - in response to Message 973304.

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
OK, from Albuquerque, NM, during peak about four hours after the Cricket graphs rose to 90+ Mbits/sec
16/ 100 = 16% | 12 107ms 18/ 100 = 18% 2/ 100 = 2% 208.68.243.254 0/ 100 = 0% | 13 107ms 16/ 100 = 16% 0/ 100 = 0% setiboincdata.ssl.berkeley.edu [208.68.240.16]

looks somewhat higher on this snapshot than during the outage.
____________

Profile RottenMutt
Avatar
Send message
Joined: 15 Mar 01
Posts: 992
Credit: 207,654,737
RAC: 0
United States
Message 973416 - Posted: 23 Feb 2010, 3:52:49 UTC - in response to Message 973304.

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
b) Tomorrow, while things are quiet during mainteance

Be good to get some 'normal case' reports over the same paths as the existing tests, to predict whether the technique will be any use during the next outage.


I hope the put off the backup or just cancel it all togather this week. they have postponed a backup a day before, and that was a shorter outage.
____________

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 973459 - Posted: 23 Feb 2010, 7:29:28 UTC - in response to Message 973416.

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
b) Tomorrow, while things are quiet during mainteance

Be good to get some 'normal case' reports over the same paths as the existing tests, to predict whether the technique will be any use during the next outage.


I hope the put off the backup or just cancel it all togather this week. they have postponed a backup a day before, and that was a shorter outage.

The last time they skipped it things were pretty ugly by the following week.
____________

Profile Peter Moss
Avatar
Send message
Joined: 15 Nov 99
Posts: 14
Credit: 1,462,882
RAC: 858
United Kingdom
Message 973474 - Posted: 23 Feb 2010, 8:50:16 UTC - in response to Message 973459.

I have two 'zones' running SETI, at home and office
My office has coped pretty well during this 'outage' and has pretty much been fully up to date UL/DL's, none pending even when I came to work monday.
However my home system (PC better than in the office), which can run SETI (CUDA) still has 60 or more results stuck in the final phase of reporting - i.e. upload completed just not reported and does not appear to have since the outage. That's all of the WU's with no more coming in.

So the routes peoples systems have to take to reach SETI may be the issue after all?

At least the other projects are 100% ok, QMC, Rosetta, Einstein and Milkyway. :)


____________

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8491
Credit: 49,780,825
RAC: 53,944
United Kingdom
Message 973482 - Posted: 23 Feb 2010, 9:59:18 UTC - in response to Message 973474.

I have two 'zones' running SETI, at home and office
My office has coped pretty well during this 'outage' and has pretty much been fully up to date UL/DL's, none pending even when I came to work monday.
However my home system (PC better than in the office), which can run SETI (CUDA) still has 60 or more results stuck in the final phase of reporting - i.e. upload completed just not reported and does not appear to have since the outage. That's all of the WU's with no more coming in.

So the routes peoples systems have to take to reach SETI may be the issue after all?

At least the other projects are 100% ok, QMC, Rosetta, Einstein and Milkyway. :)

No - I think you'll find that your home machine hasn't even tested the network to SETI recently.

This sounds like one of the recent features in BOINC has kicked in: if network congestion were really to have been the problem, there would be no point in adding to that congestion by continually retrying. The trouble is, no-one has thought of a way of getting the message "It's OK, the congested has cleared, you can retry now" through to a client which isn't retrying....

Just for your own information, have a look at the Messages tab in BOINC Manager to see if the home machine has even tried recently: and then go to the Projects tab, select SETI@home, and click the Properties button. I expect you'll see some high values for backoff times.

Just click 'Update' once. That'll start BOINC communicating again: now that the backlog has cleared, your finished tasks should be reported at the first attempt, and you should get new work within a few minutes - there's not much available at the moment, but new work is being created continuously.

archae86
Send message
Joined: 31 Aug 99
Posts: 888
Credit: 1,572,688
RAC: 5
United States
Message 973508 - Posted: 23 Feb 2010, 13:09:11 UTC - in response to Message 973351.

Have any of the path-pingers retried their packet loss counters

a) While SETI is running at maximum throughput
OK, from Albuquerque, NM, during peak about four hours after the Cricket graphs rose to 90+ Mbits/sec
16/ 100 = 16% | 12 107ms 18/ 100 = 18% 2/ 100 = 2% 208.68.243.254 0/ 100 = 0% | 13 107ms 16/ 100 = 16% 0/ 100 = 0% setiboincdata.ssl.berkeley.edu [208.68.240.16]

looks somewhat higher on this snapshot than during the outage.

Measuring again well within the first hour after the cricket graph dropped well down from the long high stable value:
0/ 100 = 0% | 12 54ms 0/ 100 = 0% 0/ 100 = 0% 208.68.243.254 1/ 100 = 1% | 13 53ms 1/ 100 = 1% 0/ 100 = 0% setiboincdata.ssl.berkeley.edu [208.68.240.16]


so it may be that in some sense this signature is one of overload--but not of the primary problem of recent times.

____________

Ranger
Send message
Joined: 27 Dec 00
Posts: 49
Credit: 39,318,995
RAC: 7,393
United States
Message 973578 - Posted: 23 Feb 2010, 17:22:43 UTC

The Cricket graph shows that Sunday night, 2010-02-21 from 22:00 to 22:30 PST there was very little activity on the network. The Berkeley router was still dropping packets at about 6%. Today, 2010-02-23 at 5:20 PST Cricket shows activity at 30+ yet there was 0% packet loss.

Running pathping on a business in the same area who also happens to be a Hurricane Electric customer yields 0% packet loss.

Do we have a faulty router at Berkeley?

____________

Profile Fred J. Verster
Volunteer tester
Send message
Joined: 21 Apr 04
Posts: 3239
Credit: 31,762,719
RAC: 4,551
Netherlands
Message 973583 - Posted: 23 Feb 2010, 17:38:01 UTC - in response to Message 973578.

Did run 2 pathpings, one from my laptop, which showed no difference.
Also one from X64 host, same results, in packet loss.
Although this host asked, after uploading results, 1 task and received 15.
This host, contacts SETI, atleast a hundred times a day and this is my only CUDA-host.
I'll watch, if this CUDA-host, does contact more times compaired to the non-CUDA-host's.

____________

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 973584 - Posted: 23 Feb 2010, 17:39:01 UTC - in response to Message 973578.

The Cricket graph shows that Sunday night, 2010-02-21 from 22:00 to 22:30 PST there was very little activity on the network. The Berkeley router was still dropping packets at about 6%. Today, 2010-02-23 at 5:20 PST Cricket shows activity at 30+ yet there was 0% packet loss.

Running pathping on a business in the same area who also happens to be a Hurricane Electric customer yields 0% packet loss.

Do we have a faulty router at Berkeley?

Either that, or the person who administers one router has pings limited, and the person who administers the other does not.
____________

NoizBoi
Send message
Joined: 30 Jan 01
Posts: 10
Credit: 3,698,476
RAC: 1,919
United Kingdom
Message 973856 - Posted: 25 Feb 2010, 21:14:43 UTC

Hey, Guess what...

In spite of a lot of people wasting their time complaining about uploads/
downloads not working, inefficiency of an under-funded system, etc ad finitum
the SETI crew sorted the problem! how many moaners contributed to this effort?
How many SETI staff complained?

If you have time to criticise others then you have too much time on your hands, use that energy to improve your own life and satisfaction.

I do not mean to offend, just state the obvious truth.

Perhaps, in future, people will be a bit more patient in this ever more complex world.......

END OF RANT!

____________

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 973866 - Posted: 25 Feb 2010, 21:37:24 UTC - in response to Message 973856.

Hey, Guess what...

In spite of a lot of people wasting their time complaining about uploads/
downloads not working, inefficiency of an under-funded system, etc ad finitum
the SETI crew sorted the problem! how many moaners contributed to this effort?
How many SETI staff complained?

If you have time to criticise others then you have too much time on your hands, use that energy to improve your own life and satisfaction.

I do not mean to offend, just state the obvious truth.

Perhaps, in future, people will be a bit more patient in this ever more complex world.......

END OF RANT!

You left out the part about staff with 8-5 monday-friday jobs who were clearly working on issues on Saturday and Sunday.
____________

Profile Sharpshooter
Send message
Joined: 26 Mar 00
Posts: 27
Credit: 2,309,165
RAC: 1,126
United States
Message 973910 - Posted: 26 Feb 2010, 1:23:02 UTC - in response to Message 973866.

Rather than complain out there get up off of a few dollars, euros, pounds or whatever and help these folks out who day in and day out bust their chops for us. We can't be there to help them but our money can. So let's get some green stars up there and show Eric and the others that we are truly in there with them. I can think of little else that could be more help to them. S.E.T.I cannot live by computer power alone...it must have financial help as well.
____________

NoizBoi
Send message
Joined: 30 Jan 01
Posts: 10
Credit: 3,698,476
RAC: 1,919
United Kingdom
Message 973929 - Posted: 26 Feb 2010, 2:52:31 UTC

Hi Ned,

I accept your notice of my omission and agree. But Im sure they were also working past 5pm!

These guys are dedicated and I take my hat off to them.

I also praise you, Ned, you only offer support and advice within the community (without getting involved in bickering about connection problems, funding, etc).

Phil H.

____________

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 973952 - Posted: 26 Feb 2010, 5:23:20 UTC - in response to Message 973929.

Hi Ned,

I accept your notice of my omission and agree. But Im sure they were also working past 5pm!

These guys are dedicated and I take my hat off to them.

I also praise you, Ned, you only offer support and advice within the community (without getting involved in bickering about connection problems, funding, etc).

Phil H.

Thanks for the nice compliment.

No matter how much bickering there is, we have to live with what's possible.

Short staff, not much funding for new stuff, and the connection is what it is.

So, BOINC is designed to tolerate problems -- and where possible, mitigate them.

... and for the most part, it works well. If you can sit back and watch when you're sure it's doing the wrong thing, you might even see that it works well.

While we don't see Matt, or Eric, or Jeff saying "we're working on it" at 11:00pm on a Saturday, we can often tell that something is being done.
____________

Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 · Next

Message boards : Technical News : Out of the fire and into the pit of sulfuric acid. (Feb 19, 2010)

Copyright © 2014 University of California