Anyone else having trouble uploading results?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
J

JWMiddleton

Diamond Member
Aug 10, 2000
5,686
172
106
torque2k, Post the method and options you are using to get to OK's proxy. It could be something you are doing wrong.

Can you get to him with this link? If so, then there is no reason why a program can't. The only except could be if it is your first try and the passthrough isn't working due to the restrictions at Berkley. To get around that I copied a numbered directory from a machine that had a cache using SetiDriver to a new machine and restarted SetiDriver on both machines. Both boxes were happy and crunchin' started! :)

 
R

Russ

Lifer
Oct 9, 1999
21,093
3
0
torque2k,

Feel free to test one of the machines by pointing it to 207.108.218.137:5517. That's the IP address of my Q, and it's static. If you can connect to it, this would indicate you're running in to a DNS issue somewhere. Then you can test at seti.tacube.com:5517. If that works, you're welcome to use it.

Russ, NCNE
 
B

BadThad

Lifer
Feb 22, 2000
12,100
49
91
I couldn't flush last night at all from home...from 6PM to 12PM EST...NO CONNECT! My home herd is getting constipated...hope I get em out tonight. :|
 
J

JWMiddleton

Diamond Member
Aug 10, 2000
5,686
172
106


<< I couldn't flush last night at all from home...from 6PM to 12PM EST... >>


They would have "opened the flood gates" at 12PM EST (9PM PST)...when it first opens I bet it is really busy!
 
L

Logix

Diamond Member
Jul 26, 2001
3,627
0
0
Yesterday, I noticed that OK was getting backed up throughout the day, until there were over 1000 work units waiting around midnight PST. Then, around 2AM things magically improved and OK was able to dump the queue. It's been mildly slow today... a wait of only about 100 WU's at OK's proxy.
 
T

torque2k

Member
Mar 1, 2001
78
0
0
JW:

<< Post the method and options you are using to get to OK's proxy. It could be something you are doing wrong. >>



I'm using the same SETI Driver settings I've been using for two weeks, orangekid.teamanandtech.com, port 5001, use colon. Worked today, though, so I know it was the queue, and not my settings or firewall ports... Thanks, though!

Russ:

<< Feel free to test one of the machines by pointing it to 207.108.218.137:5517. That's the IP address of my Q, and it's static. If you can connect to it, this would indicate you're running in to a DNS issue somewhere. Then you can test at seti.tacube.com:5517. If that works, you're welcome to use it. >>



Thanks, Russ. I'll keep your tacube address handy in case this happens again. Today, I tried the proxy for OK, and everything went through this time. I was able to get to Berkeley direct yesterday, so I at least had WUs. ;)

Musta been the load yesterday. Looks like OK's cleared up, so we're off and crunching again!

Torque2k
 
M

micron

Diamond Member
Oct 9, 1999
7,228
0
0
Maybe Berkeley could open up a "special" connection with "very high" priority for OK's queue? ;);)
 
P

Poof

Diamond Member
Jul 27, 2000
4,305
0
0
I have to agree with what JWMiddleton has said. Many of you are relative newbies with SETI and the hope is that you don't get so frustrated that you leave the project because of what is happening right now. Many of us have weathered much worse than this, which is why we preach caching so much.

In this case, since we've been told about the narrow window of opportunity (faster bandwidth) to upload/download results, then we need to set a long term strategy to deal with that by time-shifting the uploads via Setiqueue. In essence, each of these queues will act as a "mirror" server to SETI. Even if you run Setidriver (which I do along with Setiqueue), you can still point Setidriver to someone else's Setiqueue to get more WUs. Yes it is a hassle to go around to your farm to do this, but I know that the standard SETI FAQs here have always indicated that having access to some caching program is almost a must... and now you see why!

There is one possible danger to all of this though, and that's the fact that other large teams are starting to do the same type of time-shifting. So there may be times, even overnight PST, where things start to slow down as all the queues around the world start to flush. However, at least they're flushing at this time.

Unfortunately, I have a suspicion that another client that doubles the WU times might end up in the works if this continues for any extended period of time. :(
 
J

JWMiddleton

Diamond Member
Aug 10, 2000
5,686
172
106
Unfortunately, I have a suspicion that another client that doubles the WU times might end up in the works if this continues for any extended period of time. :(

I've had that same fear! Miss Poof! There is no better way to reduce their bandwidth problems in a hurry!
 
C

conjur

No Lifer
Jun 7, 2001
58,686
3
0
Maybe if we sent Ray and his cane over to the network admins at Berkeley...........

:D
 
Confused

Confused

Elite Member
Nov 13, 2000
14,166
0
0
I also have my Queue open to the public, teamanandtech.d2g.com port 5001

This is permenantly trying to connect to Berkeley, so if you wish to flush through my proxy, it will send WUs when it can, and for now i will transfer any WUs across if anyone needs them :)

ConfusedBW
 
A

aphoenixrises

Senior member
Jan 9, 2002
365
0
0
I have not had this many problems except for the past two nights. Right now I cant connect to anyones queue to get any more WU's. I have been idle for half a day. It is very frustrating. But I am still trying.:disgust:
 
K

KouklatheCat

Golden Member
Oct 23, 2000
1,502
0
0
Same here I am rarely able to connect. It is frustrating me. A clock cycle is a terrible thing to waste. My work machine sits idle from 1600 hrs till 0700 hrs. My home machine is idle from 0630 hrs till 1630 hrs with a about an hour or so of use in there somewhere. My wifes machine has been idle for a week. And my guest machine sits idle for about 20 hrs a day.
 
E

Eponymous

Golden Member
Jun 7, 2001
1,186
0
0
Maybe we should just burn our results onto a CD and mail them to Berkeley. ;)

Do we know anyone on Berkeley campus that could act as our Queue? :D
(Just to get around the bottle neck).

 
E

Eponymous

Golden Member
Jun 7, 2001
1,186
0
0
99 WU's stuck in the Queue.

99 WU's Stuck!

Take one down and pass it around...

98 Wu's Stuck in the Queue. ;)


(Ok it is late and I should get some sleep).

rolleye.gif

 
Smoke

Smoke

Distributed Computing Elite Member
Jan 3, 2001
12,650
207
106
Maybe we should just burn our results onto a CD and mail them to Berkeley.

LOL @ Grant :D

Good idea! :Q
 
M

micron

Diamond Member
Oct 9, 1999
7,228
0
0


<< Maybe we should just burn our results onto a CD and mail them to Berkeley. ;)

Do we know anyone on Berkeley campus that could act as our Queue? :D
(Just to get around the bottle neck). >>

I think Logix goes to Berkeley, let's send the CDR's to him! ;)
 
P

Poof

Diamond Member
Jul 27, 2000
4,305
0
0
aphoenixrising - what kind of errors are you getting trying to connect to the TeAm queues? If you're having probs getting WUs from someone other than SETI, that might be something on your end... :(

Also see the below suggestion to Koukla...

KouklatheCat - are your machines running windoze? If they are, you could try installing Setidriver (if you haven't already) and then point Setidriver to a TeAm Setiqueue. That way, once you get some WUs locally to crunch, if you have issues hitting the TeAm queues at some point, your machines won't sit idle for so long.
 
M

micron

Diamond Member
Oct 9, 1999
7,228
0
0
It's 7:52AM CST and OK's proxy is already filling up w/ results! :| This is getting worse every day, why doesn't SETI just have mirrors? :confused:
 
P

Poof

Diamond Member
Jul 27, 2000
4,305
0
0
It's 7:52AM CST and OK's proxy is already filling up w/ results! This is getting worse every day, why doesn't SETI just have mirrors? :confused:

Well... OK may be uploading at a specific time range overnight and then has the queue hold further results until 24 hours later. I don't think he has it set for pass-through.

And the mirror server issue was argued to the powers-that-be including Eric Kopela and they ignored it. I suspect they may have had some concerns about data integrity. But if they had been willing to negotiate agreements with some of the larger Universities world-wide, they could have possibly worked out a way to implement a secure, dedicated pipe between the various mirror servers, in order to ensure the integrity.

Oh well... looks like another doubling-of-WU-times client may hit us even earlier than I expect.
rolleye.gif
 
M

micron

Diamond Member
Oct 9, 1999
7,228
0
0


<< Oh well... looks like another doubling-of-WU-times client may hit us even earlier than I expect.
rolleye.gif
>>

I think I'll move to another project if they double WU times :frown:.
 
P

Poof

Diamond Member
Jul 27, 2000
4,305
0
0
think I'll move to another project if they double WU times . :disgust:

Awwww.... Don't leave. :)

Staying will give you the opportunity to make them look like idiots by your eventual purchase of a faster processor. Thus proving, like we're doing right now, that quick fixes giving/offering very little beneficial science in order to remedy a bandwidth problem, is only going to backfire down the road. ;)

[EDIT: I'm gonna try to check out alt.sci.seti to see what the team-less "masses" may be yelling about regarding what's going on... :p]
 
Smoke

Smoke

Distributed Computing Elite Member
Jan 3, 2001
12,650
207
106
I'm gonna try to check out alt.sci.seti to see what the team-less "masses" may be yelling about regarding what's going on... ]

Poof, this sounds like a recruiting opportunity to me. :D

Or is that a place I should not venture? :(
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom