Stats haywire? No export?


log in

Advanced search

Message boards : Number crunching : Stats haywire? No export?

Author Message
Profile Dirk Broer
Volunteer tester
Avatar
Send message
Joined: 18 Jun 00
Posts: 14
Credit: 3,052,466
RAC: 148
British Virgin Islands
Message 1171497 - Posted: 16 Nov 2011, 21:45:08 UTC

Why is it that I know to have 1.053.866 points while only 1.049.950 show?
____________

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 3963
Credit: 31,829,930
RAC: 10,488
United Kingdom
Message 1171501 - Posted: 16 Nov 2011, 21:48:15 UTC - in response to Message 1171497.

Why is it that I know to have 1.053.866 points while only 1.049.950 show?

The 1,049,950 is from an External Stats site that gets a daily stats dump, so will always be slightly out of date,

Claggy
____________


Profile Dirk Broer
Volunteer tester
Avatar
Send message
Joined: 18 Jun 00
Posts: 14
Credit: 3,052,466
RAC: 148
British Virgin Islands
Message 1171552 - Posted: 17 Nov 2011, 1:06:53 UTC - in response to Message 1171501.

Yeah, but it was showing the same value for 4 days in a row, while I kept adding results and my score kept increasing.
____________

WinterKnight
Volunteer tester
Send message
Joined: 18 May 99
Posts: 8218
Credit: 21,776,185
RAC: 12,992
United Kingdom
Message 1171586 - Posted: 17 Nov 2011, 5:17:05 UTC
Last modified: 17 Nov 2011, 5:17:31 UTC

You can see from this weekly cricket graph (green line) that the last stats output was Monday evening.

Richard Haselgrove
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8275
Credit: 44,920,306
RAC: 13,543
United Kingdom
Message 1171613 - Posted: 17 Nov 2011, 9:29:16 UTC - in response to Message 1171586.

You can see from this weekly cricket graph (green line) that the last stats output was Monday evening.

???

I don't see (well, I've never found) a way of deducing that information from Cricket.

You can always see the most recent stats dump available in http://setiathome.berkeley.edu/stats/. The timestamps are in local (Berkeley) time, so the 'current' figures are about 12 hours old. Normally, those files are regnerated once every 24 hours, though looking at BOINCstats, it appears that they may have missed on 9 and 15 November.

All SETI does is make files available in the directory I linked. Whether signatures are updated from those dumped files, and when, is entirely up to the external statistics site in question.

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12127
Credit: 2,519,625
RAC: 353
Netherlands
Message 1171622 - Posted: 17 Nov 2011, 11:55:43 UTC - in response to Message 1171613.

I'd think that since the stats are on the same page as the forums are, that they'd use the same connection to be downloaded. Thus not the one in the Cricket graph, but the other one that the whole web site runs through.
____________
Jord

Loving awareness is free.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3567
Credit: 97,843,816
RAC: 78,448
United States
Message 1171629 - Posted: 17 Nov 2011, 12:40:50 UTC - in response to Message 1171622.

I'd think that since the stats are on the same page as the forums are, that they'd use the same connection to be downloaded. Thus not the one in the Cricket graph, but the other one that the whole web site runs through.


http://setiathome.berkeley.edu is on the campus connection. Their IS/IT department does have other graphs for the rest of the campus. However IIRC everyone in the building they are in all share one connection. So that graph would not indicate exactly what you might want.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12127
Credit: 2,519,625
RAC: 353
Netherlands
Message 1171645 - Posted: 17 Nov 2011, 13:40:12 UTC - in response to Message 1171629.
Last modified: 17 Nov 2011, 13:40:24 UTC

However IIRC everyone in the building they are in all share one connection.

Not everyone. The BOINC domain (and server) runs from that building as well, even from the same server cabinet. However, it runs on a totally different internet connection, one independent from the labs and the 1Gbit connection down the hill.
____________
Jord

Loving awareness is free.

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 3963
Credit: 31,829,930
RAC: 10,488
United Kingdom
Message 1171657 - Posted: 17 Nov 2011, 14:26:20 UTC - in response to Message 1171645.

However IIRC everyone in the building they are in all share one connection.

Not everyone. The BOINC domain (and server) runs from that building as well, even from the same server cabinet. However, it runs on a totally different internet connection, one independent from the labs and the 1Gbit connection down the hill.

Are you sure?, tracert of boinc.berkeley.edu and setiathome.berkeley.edu trace the same route across campus:

C:\Users\Claggy>tracert boinc.berkeley.edu

Tracing route to boinc.berkeley.edu [128.32.18.189]
over a maximum of 30 hops:

1 127 ms 97 ms 121 ms 172.30.132.161
2 101 ms 99 ms 110 ms te-4-2.car1.Manchester1.Level3.net [195.50.119.9]
3 114 ms 116 ms 87 ms ae-4-4.ebr1.London1.Level3.net [4.69.133.102]
4 99 ms 131 ms 188 ms vlan104.ebr2.London1.Level3.net [4.69.143.98]
5 163 ms 169 ms 160 ms ae-44-44.ebr1.NewYork1.Level3.net [4.69.137.78]
6 237 ms 229 ms 172 ms ae-91-91.csw4.NewYork1.Level3.net [4.69.134.78]
7 272 ms 169 ms 200 ms ae-92-92.ebr2.NewYork1.Level3.net [4.69.148.45]
8 313 ms 286 ms 219 ms ae-2-2.ebr4.SanJose1.Level3.net [4.69.135.185]
9 240 ms 259 ms 279 ms ae-91-91.csw4.SanJose1.Level3.net [4.69.153.14]
10 344 ms 298 ms 249 ms ae-41-90.car1.SanJose1.Level3.net [4.69.152.195]
11 276 ms 262 ms 297 ms CENIC.car1.SanJose1.Level3.net [4.53.16.186]
12 230 ms 261 ms 268 ms dc-svl-core1--svl-isp1-10ge.cenic.net [137.164.47.133]
13 274 ms 279 ms 252 ms dc-svl-agg1--svl-core1-10ge.cenic.net [137.164.47.120]
14 265 ms 257 ms 251 ms dc-ucb--svl-agg1-10ge.cenic.net [137.164.50.19]
15 275 ms 260 ms 250 ms t1-3.inr-201-sut.Berkeley.EDU [128.32.0.65]
16 241 ms 228 ms 239 ms g5-1.inr-204-sut.Berkeley.EDU [128.32.255.105]
17 * * * Request timed out.
18 272 ms 310 ms 299 ms isaac.ssl.berkeley.edu [128.32.18.189]

Trace complete.

C:\Users\Claggy>tracert setiathome.berkeley.edu

Tracing route to setiathome.berkeley.edu [128.32.18.150]
over a maximum of 30 hops:

1 95 ms 100 ms 110 ms 172.30.132.161
2 233 ms 180 ms 158 ms te-4-2.car1.Manchester1.Level3.net [195.50.119.9]
3 260 ms 142 ms 150 ms ae-4-4.ebr1.London1.Level3.net [4.69.133.102]
4 166 ms 129 ms 119 ms vlan101.ebr2.London1.Level3.net [4.69.143.86]
5 246 ms 180 ms 176 ms ae-42-42.ebr1.NewYork1.Level3.net [4.69.137.70]
6 171 ms 180 ms 200 ms ae-81-81.csw3.NewYork1.Level3.net [4.69.134.74]
7 190 ms 209 ms 239 ms ae-82-82.ebr2.NewYork1.Level3.net [4.69.148.41]
8 220 ms 250 ms 349 ms ae-2-2.ebr4.SanJose1.Level3.net [4.69.135.185]
9 230 ms 239 ms 249 ms ae-61-61.csw1.SanJose1.Level3.net [4.69.153.2]
10 279 ms 260 ms 220 ms ae-11-60.car1.SanJose1.Level3.net [4.69.152.3]
11 293 ms 301 ms 237 ms CENIC.car1.SanJose1.Level3.net [4.53.16.186]
12 231 ms 219 ms 252 ms dc-svl-core1--svl-isp1-10ge.cenic.net [137.164.47.133]
13 262 ms 225 ms 260 ms dc-svl-agg1--svl-core1-10ge.cenic.net [137.164.47.120]
14 263 ms 239 ms 270 ms dc-ucb--svl-agg1-10ge.cenic.net [137.164.50.19]
15 250 ms 240 ms 338 ms t1-3.inr-201-sut.Berkeley.EDU [128.32.0.65]
16 308 ms 379 ms 250 ms g5-1.inr-204-sut.Berkeley.EDU [128.32.255.105]
17 * * * Request timed out.
18 338 ms 289 ms 446 ms thinman.ssl.berkeley.edu [128.32.18.150]

Trace complete.

Which are: Graphs for xe-1_2_0

then: Graphs for tengigabitethernet1_3

then: Graphs for gigabitethernet5_1

Then a router where the trace times out.

Claggy

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12127
Credit: 2,519,625
RAC: 353
Netherlands
Message 1171668 - Posted: 17 Nov 2011, 16:27:42 UTC - in response to Message 1171657.

Are you sure?

It's what I was always told by David, so I believe him.
That your trace goes to the same place is just because they're in the same place, end up at the same computers, the same domain (as you've noticed, both are 128.32.18.xxx). But that doesn't mean that they need to use the same outward connection. You're tracing the IP address, not the actual connection it's using.

It's happened enough times that Seti's pages were unavailable, the Cricket graphs showed minimal in- and output, while the BOINC pages were normally reachable; and vice versa, I might add. ;-)
____________
Jord

Loving awareness is free.

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12127
Credit: 2,519,625
RAC: 353
Netherlands
Message 1171686 - Posted: 17 Nov 2011, 18:10:55 UTC

OK, I did some research, here's what's happening.
Seti uses two connections, one is the 1Gbit down the hill of which they're allowed to use 100Mbit. The other is the SSL lab wide connection that the home page and forums run off of.

It's this connection that BOINC uses as well, however, the photo that purportedly shows the BOINC server (5th photo down), is out of date and no longer true.

The BOINC server runs from David's server cabinet, in his own office. It does use the SSL lab wide internet connection, but doesn't do so through the S@H switches and routers. BOINC pays for the use of office space and the like, the internet connection is included in that.

This is why Seti can be completely down, e.g. due to the AC going on strike, unpingable, untraceable, dead to the world, Cricket graphs showing no sign of live, while BOINC is up & running.
____________
Jord

Loving awareness is free.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3567
Credit: 97,843,816
RAC: 78,448
United States
Message 1171718 - Posted: 17 Nov 2011, 20:58:46 UTC - in response to Message 1171686.

OK, I did some research, here's what's happening.
Seti uses two connections, one is the 1Gbit down the hill of which they're allowed to use 100Mbit. The other is the SSL lab wide connection that the home page and forums run off of.

Odd I seem to recall reading that earlier. :)
The guys in the lab have gone over how things are setup several times over the years. Most recently the SSL lab/building got upgraded to 1GB. Maybe 6-12 months ago. I want to say it was a February, but I have slept since then. This however had nothing to do with the HE connection we use for data transfer.
It's this connection that BOINC uses as well, however, the photo that purportedly shows the BOINC server (5th photo down), is out of date and no longer true.

The BOINC server runs from David's server cabinet, in his own office. It does use the SSL lab wide internet connection, but doesn't do so through the S@H switches and routers. BOINC pays for the use of office space and the like, the internet connection is included in that.

This is why Seti can be completely down, e.g. due to the AC going on strike, unpingable, untraceable, dead to the world, Cricket graphs showing no sign of live, while BOINC is up & running.

Normally when the power goes out we loose the web servers as well if it is for an extended time, or when the UPS just freaks out and shuts everything off. However, I don't think that has happened in a long time.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

N9JFE
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 9306
Credit: 11,902,561
RAC: 14,845
United States
Message 1171970 - Posted: 18 Nov 2011, 18:58:55 UTC - in response to Message 1171718.

OK, I did some research, here's what's happening.
Seti uses two connections, one is the 1Gbit down the hill of which they're allowed to use 100Mbit. The other is the SSL lab wide connection that the home page and forums run off of.

Odd I seem to recall reading that earlier. :)
The guys in the lab have gone over how things are setup several times over the years. Most recently the SSL lab/building got upgraded to 1GB. Maybe 6-12 months ago. I want to say it was a February, but I have slept since then. This however had nothing to do with the HE connection we use for data transfer.

Okay, I'm more confused now than I was before. I know S@H has two internet connections, the UCB campus and Hurricane Electric. Which one is called "the gigabit down the hill"? Which one is limited to 100mb? What specific traffic moves over each? Can someone point me to the last time someone who actually works there updated all this? I tried following links from the S@H home page, but they still take you to how Seti Classic worked (in the present tense -- "the screen saver does [this and that]," "downloading a work unit will take about five minutes on a standard modem").

It's this connection that BOINC uses as well, however, the photo that purportedly shows the BOINC server (5th photo down), is out of date and no longer true.

The BOINC server runs from David's server cabinet, in his own office. It does use the SSL lab wide internet connection, but doesn't do so through the S@H switches and routers. BOINC pays for the use of office space and the like, the internet connection is included in that.


I know the whole reason they got the HE connection was because S@H was tying up the bandwidth for the entire campus. I also know that the campus has since upgraded, but internal politics are preventing S@H from getting a piece of that. How does BOINC fit into that?

Does the university (or anyone else) profit in any way from BOINC? Does every BOINC project depend on the BOINC server(s) at UCB for some level of administration?
____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 13,664,580
RAC: 12,096
United States
Message 1171978 - Posted: 18 Nov 2011, 19:46:15 UTC - in response to Message 1171970.

I will try to answer some of this from what little I know. The UCB connection is the 1GB connection "down the hill". SETI uses it for the main webpage and a few other things. They are limited as to how much of this they can use because the whole SSL is dependent on this line.

The HE link is also 1GB but, due to campus policy, must be run through the campus IT department down the hill. In order for SETI to get the full advantage to this link a second fiber line would have to be run up the hill exclusively for SETI. To do this would cost tens of thousands of dollars and must be approved by UCB. Until such time as they can get both the money and approval together, they are stuck on a regular copper line that only supports a max of 100MB. It is through this line that we are sending and receiving work and most everything else.

BOINC receives it's own funding and pays UCB for it's office space and internet connection. It has nothing to do with SETI as far as that. It's server is on a completely different connection. Much like SETI, the other projects are independent of BOINC and have their own admins. I believe, but don't know for sure, that they pay BOINC for using it to run their projects.

That's my understanding of the situation anyway. I'm sure someone with a better outlook will come along soon to help explain it better than I can.
____________


PROUD MEMBER OF Team Starfire World BOINC

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3567
Credit: 97,843,816
RAC: 78,448
United States
Message 1171979 - Posted: 18 Nov 2011, 20:04:08 UTC - in response to Message 1171978.

I will try to answer some of this from what little I know. The UCB connection is the 1GB connection "down the hill". SETI uses it for the main webpage and a few other things. They are limited as to how much of this they can use because the whole SSL is dependent on this line.

The HE link is also 1GB but, due to campus policy, must be run through the campus IT department down the hill. In order for SETI to get the full advantage to this link a second fiber line would have to be run up the hill exclusively for SETI. To do this would cost tens of thousands of dollars and must be approved by UCB. Until such time as they can get both the money and approval together, they are stuck on a regular copper line that only supports a max of 100MB. It is through this line that we are sending and receiving work and most everything else.

BOINC receives it's own funding and pays UCB for it's office space and internet connection. It has nothing to do with SETI as far as that. It's server is on a completely different connection. Much like SETI, the other projects are independent of BOINC and have their own admins. I believe, but don't know for sure, that they pay BOINC for using it to run their projects.

That's my understanding of the situation anyway. I'm sure someone with a better outlook will come along soon to help explain it better than I can.

In one of their postings, in Technical News or News, about the connection it was mentioned that if all the political BS was overcome. That there would be some added cost of using the line at a higher speed. They didn't elaborate, but it could be something to do with monthly data usage.

So if the line, and the two routers on either end of it, get spoofed up they could limit the bandwidth on the line down to 100mb.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Wembley
Volunteer tester
Avatar
Send message
Joined: 16 Sep 09
Posts: 415
Credit: 888,257
RAC: 0
United States
Message 1172000 - Posted: 18 Nov 2011, 21:29:15 UTC - in response to Message 1171978.

The HE link is also 1GB but, due to campus policy, must be run through the campus IT department down the hill. In order for SETI to get the full advantage to this link a second fiber line would have to be run up the hill exclusively for SETI. To do this would cost tens of thousands of dollars and must be approved by UCB. Until such time as they can get both the money and approval together, they are stuck on a regular copper line that only supports a max of 100MB. It is through this line that we are sending and receiving work and most everything else.


If this is true it might be beneficial to move the download servers from the SETI lab, to a location down the hill on the other end of this 100MB link, so that work units only need to be transferred on this link once.
____________


Donate with your searches and online buys:
http://www.goodsearch.com/toolbar/university-of-california-setihome

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3567
Credit: 97,843,816
RAC: 78,448
United States
Message 1172077 - Posted: 19 Nov 2011, 2:40:06 UTC - in response to Message 1172000.

The HE link is also 1GB but, due to campus policy, must be run through the campus IT department down the hill. In order for SETI to get the full advantage to this link a second fiber line would have to be run up the hill exclusively for SETI. To do this would cost tens of thousands of dollars and must be approved by UCB. Until such time as they can get both the money and approval together, they are stuck on a regular copper line that only supports a max of 100MB. It is through this line that we are sending and receiving work and most everything else.


If this is true it might be beneficial to move the download servers from the SETI lab, to a location down the hill on the other end of this 100MB link, so that work units only need to be transferred on this link once.

I would just ask. Who is going to pay for this location & the fibre channel link to the storage array in the lab? Along with them having to go to 2 locations to admin their boxes.

Most of what I have heard is even at 100mb there are more connections than can be served. So going to a full gb line would not really help that much right now.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Message boards : Number crunching : Stats haywire? No export?

Copyright © 2014 University of California