LLRnet servers for NPLB - Page 13

IronBits · 2008-07-10, 06:55

I stopped each server and ran
./llrnet llrserver.lua -s three times
restarted each server
port 10,000 knpairs.txt shows
first k/n pair 147 600830

I just looked at joblist.txt for port 10000, it shows abandoned... I assume at some point, those should show backup in knpairs.txt ?
I also assume those should show up in the correct order so the smallest knpair is at the top?

jobList = {
["147/600830"] = {
["seconds"]=1213807622,
["user"]="Anonymous",
["date"]="18/06/2008 09:47:02 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:26 AM",
["result"]="CANCEL",
["n"]="600830",
},
["147/600841"] = {
["seconds"]=1213807703,
["user"]="Anonymous",
["date"]="18/06/2008 09:48:23 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:27 AM",
["result"]="CANCEL",
["n"]="600841",
},
...
["147/601004"] = {
["seconds"]=1213808743,
["user"]="Anonymous",
["date"]="18/06/2008 10:05:43 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:41 AM",
["result"]="CANCEL",
["n"]="601004",
},
["147/601005"] = {
["seconds"]=1213809278,
["user"]="Anonymous",
["date"]="18/06/2008 10:14:38 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:42 AM",
["result"]="CANCEL",
["n"]="601005",
},

IronBits · 2008-07-10, 06:59

Quote:

Originally Posted by Lennart

Try run in windoze llrserver -s
In linux ./llrnet llrserver.lua -s

See if it helps.

/Lennart

What exactly is -s telling the llrnet server to do?
We are forcing something that should be happening automatically?

Again, thanks for the tip! :)

IronBits · 2008-07-10, 07:02

Can someone direct me to the correct person to discuss these short commings so I can understand better what it should be doing?
If not, I may have to fall back to a Windows based Server.
I also need to figure out why MySQL won't work with llrnet server.lua
Running CentOS 5.1 64bit

Brucifer · 2008-07-10, 07:06

Hey IB, don't be shy!

This is an interesting thread!! At least you know enough to ask questions -- that's better than some of us, including me!

IronBits · 2008-07-10, 07:24

Thanks! I'm just getting frustrated because the software is not doing what it should be doing and is requiring a lot of babysitting, which is should not.
Documentation is very sparse, and I'm getting bits & pieces from all over the place.
Learning curve is steep when I don't know what it should or shouldn't be doing, and I can't understand a thing about what the project does, other than look for prime numbers.
Anything beyond that gives me headaches trying to understand the math behind it.
I just wanted to help you all out, and running servers is easy for me, however, this software package leaves a lot to be desired.
It was puking on the Windows based Servers so I moved it to a Linux based Server and that appeared to clear up quite a few problems, including solving the 'running out of sockets' on the Windows box, thus freezing the llrnet server process.
Then it turns out, that even under Linux (32 or 64bit) if you get more than 50 computers pounding away at it, it still locks up the llrnet server process.
/bangs head
I just want it to work smoothly, correctly and without all the babysitting, as it should be.
Someone has to know how this software works, after all, someone wrote it?!

gd_barnes · 2008-07-10, 07:44

Oh boy, Lennert or Adam, can you chime in here? Adam's port 300 is working very smoothly and as far as I can tell is correctly reporting the first k/n pair remaining. I will send a PM to Adam and refer him to this thread.

David, I'm sorry about your frustrations. I'm trying to logically analyze the way the files are working without knowing everything about them. I wasn't aware of these "abandoned" statuses.

If we can't get things clear from Adam, I'll contact Jean Penne, whom I believe was responsible for putting together a big portion of the server software, and refer him here.

Gary

em99010pepe · 2008-07-10, 08:47

Quote:

Originally Posted by IronBits

What exactly is -s telling the llrnet server to do?
We are forcing something that should be happening automatically?

Again, thanks for the tip! :)

Why didn't you ask the smartest guy on this forum...me...me...me...here you are:

Quote:

-h :
print this message

-d :
detach server and run in background

-s :
simplify joblist and knpairs files by removing solved pairs
from these files

-sort-joblist :
sort the "joblist.txt" file and write it out as
"sorted-joblist.txt" file

-import-jobs :
use this option when you upgrade your llrnet from non sql to sql.
NOTE : this also prune knpairs and joblist and import the pairs.
Be sure you have configured you sql options correctly into
llr-serverconfig.txt first.

-import-pairs :
append pairs from knpairs.txt into the sql database.

-import-results :
import results from results.txt into sql database.
date must be in this type of format : 'Sun Feb 6 00:57:42 2005' or
in mysql compatible date format (eg. '2005-2-4 23:23:12').

-import-rejected :
import rejected results from rejected.txt into sql database.
(see above for supported date formats)

-create-tables :
create the sql tables necessary for llrnet.

Lennart · 2008-07-10, 10:45

Quote:

Originally Posted by IronBits

I stopped each server and ran
./llrnet llrserver.lua -s three times
restarted each server
port 10,000 knpairs.txt shows
first k/n pair 147 600830

I just looked at joblist.txt for port 10000, it shows abandoned... I assume at some point, those should show backup in knpairs.txt ?
I also assume those should show up in the correct order so the smallest knpair is at the top?

jobList = {
["147/600830"] = {
["seconds"]=1213807622,
["user"]="Anonymous",
["date"]="18/06/2008 09:47:02 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:26 AM",
["result"]="CANCEL",
["n"]="600830",
},
["147/600841"] = {
["seconds"]=1213807703,
["user"]="Anonymous",
["date"]="18/06/2008 09:48:23 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:27 AM",
["result"]="CANCEL",
["n"]="600841",
},
...
["147/601004"] = {
["seconds"]=1213808743,
["user"]="Anonymous",
["date"]="18/06/2008 10:05:43 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:41 AM",
["result"]="CANCEL",
["n"]="601004",
},
["147/601005"] = {
["seconds"]=1213809278,
["user"]="Anonymous",
["date"]="18/06/2008 10:14:38 AM",
["k"]="147",
["status"]="abandonned",
["resultdate"]="18/06/2008 10:15:42 AM",
["result"]="CANCEL",
["n"]="601005",
},

This is what i got when i conected to port 10000

Cache is full with 1 WU.
LOGGING OUT
LLRMain started
Working on : 147/600830 (600000000000:M:1:2:258)
Starting Lucas Lehmer Riesel prime test of 147*2^600830-1
Using Irrational Base DWT : Mersenne fftlen = 32768, Used fftlen = 40960
V1 = 3 ; Computing U0...
V1 = 3 ; Computing U0...done.
Starting Lucas-Lehmer loop...
StartServiceCtrlDispatcher returns 0
WUCacheSize = 1
Refill threshold = 1

/Lennart

gd_barnes · 2008-07-10, 20:10

Here is the reply I got from Adam in a PM.

Quote:

I had many issues using the joblist and knpairs text files as well. My setup no longer uses them.

The k/n pair remaining info comes from a PHP script which queries the llrnet MySQL database.

I suggest focusing efforts on getting MySQL working. It will make it much easier to deal with.

Sorry this doesn't help much with your current setup. Obviously MySQL helped his setup greatly.

David, you know what? Don't worry about this anymore. It's not worth it. If you just pick up the first k/n pair in knpairs.txt as the first k/n pair still remaining, that's good enough. The pruning period change and purging of the "solved" k/n pairs in joblist.txt for port 400 may be all that was needed.

It's obvious by the test here by Lennert and previously by me that the k/n pairs are returned to the server if they go into this 'abandonded' status.

We've virtually never had an issue with k/n pairs going off into never-never land. If they are "abandonded", they always come back at some point and that is the important thing. The issue has only been the display on your web page of the first k/n pair remaining.

Some things just aren't worth it. This is a recreational activity so I don't want people to stress over nit-picky issues like this.

Thanks for your effort in trying to figure it out.

Gary

IronBits · 2008-07-11, 03:40

Thanks for the info. I'll try again to get MySQL working this week-end and maybe we will all be happier for it.

gd_barnes · 2008-07-11, 04:38

As you all may know from another thread, I sent the n=~506.3K-510K range to Adam for loading into port 300 on Weds. night. It hasn't been loaded yet and I haven't heard from him on it.

At our current rate, we're likely dry the server before noon CDT US on Friday (5 PM GMT).

Therefore I'm pulling 22 cores off of it and putting them on port 400, which will be the next priority since it contains n=~500.6K-505K.

Even with that, port 300 likely will dry by 5-7 AM GMT on Sat. with the current level of resources on it minus my own.

If you don't see a note from Adam or me by midnight-1 AM GMT Sat. morning that the file has been loaded, you may want to change your machines to port 400 by 3 AM to avoid having them sit idle.

Gary

2008-07-10, 06:55	#133
IronBits I ♥ BOINC! Oct 2002 Glendale, AZ. (USA) 3·7·53 Posts	I stopped each server and ran ./llrnet llrserver.lua -s three times restarted each server port 10,000 knpairs.txt shows first k/n pair 147 600830 I just looked at joblist.txt for port 10000, it shows abandoned... I assume at some point, those should show backup in knpairs.txt ? I also assume those should show up in the correct order so the smallest knpair is at the top? jobList = { ["147/600830"] = { ["seconds"]=1213807622, ["user"]="Anonymous", ["date"]="18/06/2008 09:47:02 AM", ["k"]="147", ["status"]="abandonned", ["resultdate"]="18/06/2008 10:15:26 AM", ["result"]="CANCEL", ["n"]="600830", }, ["147/600841"] = { ["seconds"]=1213807703, ["user"]="Anonymous", ["date"]="18/06/2008 09:48:23 AM", ["k"]="147", ["status"]="abandonned", ["resultdate"]="18/06/2008 10:15:27 AM", ["result"]="CANCEL", ["n"]="600841", }, ... ["147/601004"] = { ["seconds"]=1213808743, ["user"]="Anonymous", ["date"]="18/06/2008 10:05:43 AM", ["k"]="147", ["status"]="abandonned", ["resultdate"]="18/06/2008 10:15:41 AM", ["result"]="CANCEL", ["n"]="601004", }, ["147/601005"] = { ["seconds"]=1213809278, ["user"]="Anonymous", ["date"]="18/06/2008 10:14:38 AM", ["k"]="147", ["status"]="abandonned", ["resultdate"]="18/06/2008 10:15:42 AM", ["result"]="CANCEL", ["n"]="601005", }, Last fiddled with by IronBits on 2008-07-10 at 06:58

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PRPnet servers for NPLB	mdettweiler	No Prime Left Behind	228	2018-12-26 04:50
Servers for NPLB	gd_barnes	No Prime Left Behind	0	2009-08-10 19:21
LLRnet servers for CRUS	gd_barnes	Conjectures 'R Us	39	2008-07-15 10:26
NPLB LLRnet server discussion	em99010pepe	No Prime Left Behind	229	2008-04-30 19:13
NPLB LLRnet server #1 - dried	em99010pepe	No Prime Left Behind	19	2008-03-26 06:19

2008-07-10, 07:02	#135
IronBits I ♥ BOINC! Oct 2002 Glendale, AZ. (USA) 3·7·53 Posts	Can someone direct me to the correct person to discuss these short commings so I can understand better what it should be doing? If not, I may have to fall back to a Windows based Server. I also need to figure out why MySQL won't work with llrnet server.lua Running CentOS 5.1 64bit

2008-07-10, 07:06	#136
Brucifer Dec 2005 139₁₆ Posts	Hey IB, don't be shy! This is an interesting thread!! At least you know enough to ask questions -- that's better than some of us, including me!

2008-07-10, 07:24	#137
IronBits I ♥ BOINC! Oct 2002 Glendale, AZ. (USA) 3·7·53 Posts	Thanks! I'm just getting frustrated because the software is not doing what it should be doing and is requiring a lot of babysitting, which is should not. Documentation is very sparse, and I'm getting bits & pieces from all over the place. Learning curve is steep when I don't know what it should or shouldn't be doing, and I can't understand a thing about what the project does, other than look for prime numbers. Anything beyond that gives me headaches trying to understand the math behind it. I just wanted to help you all out, and running servers is easy for me, however, this software package leaves a lot to be desired. It was puking on the Windows based Servers so I moved it to a Linux based Server and that appeared to clear up quite a few problems, including solving the 'running out of sockets' on the Windows box, thus freezing the llrnet server process. Then it turns out, that even under Linux (32 or 64bit) if you get more than 50 computers pounding away at it, it still locks up the llrnet server process. /bangs head I just want it to work smoothly, correctly and without all the babysitting, as it should be. Someone has to know how this software works, after all, someone wrote it?!

2008-07-10, 07:44	#138
gd_barnes May 2007 Kansas; USA 10395₁₀ Posts	Oh boy, Lennert or Adam, can you chime in here? Adam's port 300 is working very smoothly and as far as I can tell is correctly reporting the first k/n pair remaining. I will send a PM to Adam and refer him to this thread. David, I'm sorry about your frustrations. I'm trying to logically analyze the way the files are working without knowing everything about them. I wasn't aware of these "abandoned" statuses. If we can't get things clear from Adam, I'll contact Jean Penne, whom I believe was responsible for putting together a big portion of the server software, and refer him here. Gary

2008-07-11, 03:40	#142
IronBits I ♥ BOINC! Oct 2002 Glendale, AZ. (USA) 3×7×53 Posts	Thanks for the info. I'll try again to get MySQL working this week-end and maybe we will all be happier for it.

2008-07-11, 04:38	#143
gd_barnes May 2007 Kansas; USA 3³·5·7·11 Posts	As you all may know from another thread, I sent the n=~506.3K-510K range to Adam for loading into port 300 on Weds. night. It hasn't been loaded yet and I haven't heard from him on it. At our current rate, we're likely dry the server before noon CDT US on Friday (5 PM GMT). Therefore I'm pulling 22 cores off of it and putting them on port 400, which will be the next priority since it contains n=~500.6K-505K. Even with that, port 300 likely will dry by 5-7 AM GMT on Sat. with the current level of resources on it minus my own. If you don't see a note from Adam or me by midnight-1 AM GMT Sat. morning that the file has been loaded, you may want to change your machines to port 400 by 3 AM to avoid having them sit idle. Gary