Message boards :
Number crunching :
STOP forced user update of seti projects as device gigabitethernet0_1: 66.28.250.121: cogent ssl net is dying on its feet
Message board moderation
Author | Message |
---|---|
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
Have you seen the error rate for this device. Its about to die I think. Clearly backlogs generate errors - probably collisions - and this impedes recovery from down time situations. Users should hold back from forced updates to seti I think to give the local network time to recover. Alternatively....there could be a medium/device problem. Take a look at this graph http://www.net.berkeley.edu:8885/~cricket/cricket/grapher.cgi?target=/router-interfaces/inr-668-interfaces/gigabitethernet0_1&ranges=d%3Aw&view=Errors we're approaching 370 million errors per second..... If this is a collision problem the seti team might need to think about using more sub nets from this device to reduce the problem. |
WerK Send message Joined: 30 Jun 02 Posts: 26 Credit: 221,390 RAC: 0 |
Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s |
MikeSW17 Send message Joined: 3 Apr 99 Posts: 1603 Credit: 2,700,523 RAC: 0 |
"About to Die" Ian? It's been dead so long, it's beginnig to decompose ;) |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
<blockquote>we're approaching 370 million errors per second..... If this is a collision problem the seti team might need to think about using more sub nets from this device to reduce the problem.</blockquote> m = milli, M = Mega. ;) Just look on the daily graph, there was a spike at 1.1 error/second last Tuesday. Todays spike is much lower. |
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
|
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
|
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
<blockquote>Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s</blockquote> Ah so if you look at the left of the graph you say it reads 370 milli errors per second? Duh? What sense does that make? Take another look dude.....consider context....feasibilty....and re think it. On your estimate then we have one third of an error per second......hehe...an error is always 1...you never get a part error! Oh and which "m" is micro please? |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13752 Credit: 208,696,464 RAC: 304 |
<blockquote>you never get a part error!</blockquote> Look at the Daily Graph. It has .2 per division & there are the odd spikes that fall well short of 1.0 & no m anywhere on the vertical scale. I think the units need better defining. Grant Darwin NT |
WerK Send message Joined: 30 Jun 02 Posts: 26 Credit: 221,390 RAC: 0 |
<blockquote><blockquote>Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s</blockquote> Ah so if you look at the left of the graph you say it reads 370 milli errors per second? Duh? What sense does that make? Take another look dude.....consider context....feasibilty....and re think it. On your estimate then we have one third of an error per second......hehe...an error is always 1...you never get a part error! Oh and which "m" is micro please?</blockquote> Micro is not 'm' it is µ. And please dont be so offensive, its not my fault that you were sleeping at school. If you look left where traffic in bits/s is shown, you will see that numbers are suffixed with 'M' which stands for Mega. Why errors aren't ? |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13752 Credit: 208,696,464 RAC: 304 |
Errors appear to have stopped (for now). Data rate still pegged at 90Mb/s. Grant Darwin NT |
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
|
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
<blockquote><blockquote><blockquote>Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s</blockquote> Ah so if you look at the left of the graph you say it reads 370 milli errors per second? Duh? What sense does that make? Take another look dude.....consider context....feasibilty....and re think it. On your estimate then we have one third of an error per second......hehe...an error is always 1...you never get a part error! Oh and which "m" is micro please?</blockquote> Micro is not 'm' it is µ. And please dont be so offensive, its not my fault that you were sleeping at school. If you look left where traffic in bits/s is shown, you will see that numbers are suffixed with 'M' which stands for Mega. Why errors aren't ?</blockquote> I'm sory if I upset anyone. The graph says errors per second......nothing about bits per second. The measure is millions. There is no such thing as a milli error or micro error: impossible. I use this software regularly..... Any how my message was to stop forcing an update as it exacerbates a bad problem. |
MikeSW17 Send message Joined: 3 Apr 99 Posts: 1603 Credit: 2,700,523 RAC: 0 |
Seems to me that the errors and max traffic are on the Outbound-From-Berkeley side? The servers seem to be trying to send out more traffic than the link can carry? So much for the servers being overloaded.... |
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
<blockquote>Seems to me that the errors and max traffic are on the Outbound-From-Berkeley side? The servers seem to be trying to send out more traffic than the link can carry? So much for the servers being overloaded.... </blockquote> I think...think... priority goes to downloads. So I guess the out traffic will be higher...and there is more data than an upload so your logic is right for sure. So its killing itself really! That is one hell of a busy local network right now. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13752 Credit: 208,696,464 RAC: 304 |
<blockquote>The graph says errors per second......nothing about bits per second. The measure is millions. There is no such thing as a milli error or micro error: impossible. I use this software regularly.....</blockquote> Looking at the graphs, that one is the only one with the scale as 100 m, 200 m, 300 m, 400 m. All the other error graphs have it as .5, 1.0. No m anywhere. Grant Darwin NT |
WerK Send message Joined: 30 Jun 02 Posts: 26 Credit: 221,390 RAC: 0 |
<blockquote><blockquote><blockquote><blockquote>Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s</blockquote> Ah so if you look at the left of the graph you say it reads 370 milli errors per second? Duh? What sense does that make? Take another look dude.....consider context....feasibilty....and re think it. On your estimate then we have one third of an error per second......hehe...an error is always 1...you never get a part error! Oh and which "m" is micro please?</blockquote> Micro is not 'm' it is µ. And please dont be so offensive, its not my fault that you were sleeping at school. If you look left where traffic in bits/s is shown, you will see that numbers are suffixed with 'M' which stands for Mega. Why errors aren't ?</blockquote> I'm sory if I upset anyone. The graph says errors per second......nothing about bits per second. The measure is millions. There is no such thing as a milli error or micro error: impossible. I use this software regularly..... Any how my message was to stop forcing an update as it exacerbates a bad problem.</blockquote> Yes, you are true that people must stop hitting update like no-brains because it slows everyone down. I was talking about difference between the error/sec graph ( such as this ) and the one which shows traffic in bits/sec (such as these ). My point is that if it really was millions errors/sec it would be suffixed with 'M' such as bits/sec in the traffic graph. Nothing more. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13752 Credit: 208,696,464 RAC: 304 |
<blockquote>I think...think... priority goes to downloads.</blockquote> I'll vote the opposite. After an outage i can always return (at random) Results long before i'm able to even contact the Schedulers. Even then, almost all Results will always be returned before i can start downloading the odd new Work Unit to process. Grant Darwin NT |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
<blockquote> The graph says errors per second......nothing about bits per second. The measure is millions. There is no such thing as a milli error or micro error: impossible. I use this software regularly..... Any how my message was to stop forcing an update as it exacerbates a bad problem. </blockquote> Well, if you compares with the other graphs, you'll see the bandwith-graph is using M for Megabits transferred, and on the packets sent they're using k for kilo. Also, if you take a quick look on the daily error-graph, you'll see there is no mentioning of m or M here, and the spike last Tuesday was on 1.1 error per second. Remember, the cricket-graphs is being updated every 5 minutes, 5 minutes is 300 seconds, and the rates reported is average. One error averaged over 300 seconds is... 3.3333 milli-error/second. As for not forcing an update, this is a good advice. :) |
Tigher Send message Joined: 18 Mar 04 Posts: 1547 Credit: 760,577 RAC: 0 |
<blockquote><blockquote><blockquote><blockquote><blockquote>Well, I think you are wrong. When number is suffixed with 'm' it stands for 'mili' (10^-3) , so its 370 mili errors/sec = 0.37 errors/sec . When number is suffixed with 'M' it stands for 'mega' (10^6) such as 70 Mb/s</blockquote> Ah so if you look at the left of the graph you say it reads 370 milli errors per second? Duh? What sense does that make? Take another look dude.....consider context....feasibilty....and re think it. On your estimate then we have one third of an error per second......hehe...an error is always 1...you never get a part error! Oh and which "m" is micro please?</blockquote> Micro is not 'm' it is µ. And please dont be so offensive, its not my fault that you were sleeping at school. If you look left where traffic in bits/s is shown, you will see that numbers are suffixed with 'M' which stands for Mega. Why errors aren't ?</blockquote> I'm sory if I upset anyone. The graph says errors per second......nothing about bits per second. The measure is millions. There is no such thing as a milli error or micro error: impossible. I use this software regularly..... Any how my message was to stop forcing an update as it exacerbates a bad problem.</blockquote> Yes, you are true that people must stop hitting update like no-brains because it slows everyone down. I was talking about difference between the error/sec graph ( such as <A>this[/url] ) and the one which shows traffic in bits/sec (such as <A>this[/url] ). My point is that if it really was millions errors/sec it would be suffixed with 'M' such as bits/sec in the traffic graph. Nothing more.</blockquote> Ok mate...sorry....really! I agree. The IT industry is famous for this vague nomenclature. I mean...I keep arguing with my son's teacher that bps is bits per second and Bps is bytes per second....which I think is right. She says no not these days. God help us...there's quite a difference. Ha and computing wants to be a science lol! Long way to go there. Again sorry mate meant no offence!. |
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
I've already uploaded and reported my work. It didn't ask for a download since it's version 4.37. Maybe I got mine in before all of you woke up. LOL Can't we all just get along???? Yep, there are errors, however you interpret it. If you left for 24 hours leaving things completely alone, then when you got back everything would probably be just peachy. |
©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.