PRPnet beta test server

[IronBits]

Whether the Servers are running or not has zero impact to the copying off of the files.
Servers do not need to be stopped for that to happen because, it's being run on Linux not Windows

My Servers run 24/7 unless you give me more k/n pairs to add to knpairs.txt at which time I stop, add the pairs, and then re-start the server.

[Flatlander]

I bet Gary used to jump on his toys when he was small, to see if they would break.

I agree though. After the bug with Phrot we don't want one here too.

[gd_barnes]

Originally Posted by IronBits

Whether the Servers are running or not has zero impact to the copying off of the files.
Servers do not need to be stopped for that to happen because, it's being run on Linux not Windows

My Servers run 24/7 unless you give me more k/n pairs to add to knpairs.txt at which time I stop, add the pairs, and then re-start the server.

Let me restate:

We need to see if the copying off still fully works IF the server is stopped in the middle of it's copying process for any reason whatsoever...since that can happen for any reason at any time in the real world.

I'm not saying that they need to be stopped for the copying off to work.

In other words, we want to test the server in all kinds of wierd situations that it might encounter in the real world while we are beta testing. In this case, we want to make sure that it completes its copying process if it has started but not finished it when it is stopped and then restarted. I'm more concerned about the interfacing between the PRPnet server and the stats/results process than anything. I do not want us to lose or have duplicate results because it can be such a mess to straighten out.

Funny you should mention that Chris. No, I didn't like to "stress test" my toys (lol) as a kid but as an adult and a mainframe programmer, I used to like to find ways to break programs. I'd do all kinds of things to them: Feed them empty files, put alphas in numeric fields, run huge files through them, run files through that I knew would cause monetary overflow (or underflow) in some fields, etc. When you're running a back office (i.e. accounting and billing) system for an insurance co., you better make sure you fully test new programs or changes to existing programs before putting them in the production environment.


Gary

[Razor_FX_II]

Has anyone experienced a client not saying in the window that its returning or getting new work but get new work anyway.
I restarted it and then it ran fine.
Not sure if these that ran one right after another were returned or not.
I checked the log file and here is the results.

[2009-05-03 03:46:40 GMT] User name Razor_FX_II at email address is razorfxii@sio.midco.net
[2009-05-03 03:49:05 GMT] G2000: 3099*2^428793-1 is not prime.  Residue 21E42EC297B0A5E3
[2009-05-03 03:52:28 GMT] G2000: 3123*2^428794-1 is not prime.  Residue A2129B9C15D2D9B1
[2009-05-03 03:55:52 GMT] G2000: 3145*2^428793-1 is not prime.  Residue CFE6B12A0E73F935
[2009-05-03 03:59:16 GMT] G2000: 3149*2^428796-1 is not prime.  Residue 631725BCAB5164B6
[2009-05-03 04:02:51 GMT] G2000: 3183*2^428795-1 is not prime.  Residue 48237763EB8D9A16
[2009-05-03 04:06:16 GMT] G2000: 3189*2^428796-1 is not prime.  Residue 7F9255F6C955D2F7
[2009-05-03 04:09:40 GMT] G2000: 3195*2^428793-1 is not prime.  Residue FC1A5ADB3F461158
[2009-05-03 04:12:14 GMT] G2000: Could not open file [lresults.txt] for reading.  Assuming user stopped with ^C
[2009-05-03 04:12:58 GMT] G2000: Attempting to send previously completed work
[2009-05-03 04:12:59 GMT] G2000: Returning work to server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:13:03 GMT] G2000: INFO: Test for candidate 3099*2^428793-1 accepted
[2009-05-03 04:13:03 GMT] PRPNet Client application v2.0.4 started
[2009-05-03 04:13:03 GMT] User name Razor_FX_II at email address is razorfxii@sio.midco.net
[2009-05-03 04:13:04 GMT] G2000: Getting work from server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:16:33 GMT] G2000: 3151*2^432301-1 is not prime.  Residue 9584289E3B7DF743
[2009-05-03 04:16:34 GMT] Total Time:  0:03:35  Total Tests: 1  Total PRPs Found: 0
[2009-05-03 04:16:34 GMT] G2000: Returning work to server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:16:37 GMT] G2000: INFO: Test for candidate 3151*2^432301-1 accepted
[2009-05-03 04:16:37 GMT] G2000: Getting work from server nplb-gb1.no-ip.org at port 2000

[mdettweiler]

Originally Posted by Razor_FX_II

Has anyone experienced a client not saying in the window that its returning or getting new work but get new work anyway.
I restarted it and then it ran fine.
Not sure if these that ran one right after another were returned or not.
I checked the log file and here is the results.

[2009-05-03 03:46:40 GMT] User name Razor_FX_II at email address is razorfxii@sio.midco.net
[2009-05-03 03:49:05 GMT] G2000: 3099*2^428793-1 is not prime.  Residue 21E42EC297B0A5E3
[2009-05-03 03:52:28 GMT] G2000: 3123*2^428794-1 is not prime.  Residue A2129B9C15D2D9B1
[2009-05-03 03:55:52 GMT] G2000: 3145*2^428793-1 is not prime.  Residue CFE6B12A0E73F935
[2009-05-03 03:59:16 GMT] G2000: 3149*2^428796-1 is not prime.  Residue 631725BCAB5164B6
[2009-05-03 04:02:51 GMT] G2000: 3183*2^428795-1 is not prime.  Residue 48237763EB8D9A16
[2009-05-03 04:06:16 GMT] G2000: 3189*2^428796-1 is not prime.  Residue 7F9255F6C955D2F7
[2009-05-03 04:09:40 GMT] G2000: 3195*2^428793-1 is not prime.  Residue FC1A5ADB3F461158
[2009-05-03 04:12:14 GMT] G2000: Could not open file [lresults.txt] for reading.  Assuming user stopped with ^C
[2009-05-03 04:12:58 GMT] G2000: Attempting to send previously completed work
[2009-05-03 04:12:59 GMT] G2000: Returning work to server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:13:03 GMT] G2000: INFO: Test for candidate 3099*2^428793-1 accepted
[2009-05-03 04:13:03 GMT] PRPNet Client application v2.0.4 started
[2009-05-03 04:13:03 GMT] User name Razor_FX_II at email address is razorfxii@sio.midco.net
[2009-05-03 04:13:04 GMT] G2000: Getting work from server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:16:33 GMT] G2000: 3151*2^432301-1 is not prime.  Residue 9584289E3B7DF743
[2009-05-03 04:16:34 GMT] Total Time:  0:03:35  Total Tests: 1  Total PRPs Found: 0
[2009-05-03 04:16:34 GMT] G2000: Returning work to server nplb-gb1.no-ip.org at port 2000
[2009-05-03 04:16:37 GMT] G2000: INFO: Test for candidate 3151*2^432301-1 accepted
[2009-05-03 04:16:37 GMT] G2000: Getting work from server nplb-gb1.no-ip.org at port 2000

Well, based on some discussions I've had with Mark via email, you definitely shouldn't be losing any results. Whenever the client says "Returning work to server x at port x" it always tries to return all queued results for that server. So, you should be OK.

[MyDogBuster]

Funny you should mention that Chris. No, I didn't like to "stress test" my toys (lol) as a kid but as an adult and a mainframe programmer, I used to like to find ways to break programs. I'd do all kinds of things to them: Feed them empty files, put alphas in numeric fields, run huge files through them, run files through that I knew would cause monetary overflow (or underflow) in some fields, etc. When you're running a back office (i.e. accounting and billing) system for an insurance co., you better make sure you fully test new programs or changes to existing programs before putting them in the production environment.

I can relate to Gary's testing ideas. I was a mainframe back office programmer also (before I got old and cute). The easiest way to debug or test a program is to make it fail. Feed it junk, pull the plug, run it with other programs that share the same DB. And when you think you have it perfect and put it into production, be sure to keep your evenings free for the first couple months of production runs. The phone will ring, and you will be going in to fix it anyway. Before Murphy wrote all his laws, he was a programmer. LOL

[rogue]

Originally Posted by mdettweiler

Well, based on some discussions I've had with Mark via email, you definitely shouldn't be losing any results. Whenever the client says "Returning work to server x at port x" it always tries to return all queued results for that server. So, you should be OK.

This is not an issue. The client is not capturing the other INFO messages from the server and outputting them. It will be fixed in the next release. I hope to release a beta tomorrow.

[gd_barnes]

Can someone please define the word "binary" as it is used in this thread? Before I got into DC 2 years ago, I never heard the term used in the way that it is frequently used on this forum.

It appears that it means some sort of special program but can someone be more specific? Perhaps a Linux or math-related program?


Thanks,
Gary

[Mini-Geek]

Originally Posted by gd_barnes

Can someone please define the word "binary" as it is used in this thread? Before I got into DC 2 years ago, I never heard the term used in the way that it is frequently used on this forum.

It appears that it means some sort of special program but can someone be more specific? Perhaps a Linux or math-related program?


Thanks,
Gary

In short: It means an executable, I don't know why we call it a binary.
In long:
It means a compiled program (as opposed to a program still in source code form). a.k.a. executable. (this is not a Linux- or math-specific term) On Windows this is a .exe, on other systems it's...um...not that simple...no extension, I think, but somehow the OS knows it's an executable. I don't really know precisely.
It's probably used here so much more than other places because the source code is (almost?) always available, and there is much variation in OSs and bits (e.g. Win32, Win64, Linux32, Linux64, etc. would all have the same source code but different binaries).

[mdettweiler]

Originally Posted by Mini-Geek

In short: It means an executable, I don't know why we call it a binary.
In long:
It means a compiled program (as opposed to a program still in source code form). a.k.a. executable. (this is not a Linux- or math-specific term) On Windows this is a .exe, on other systems it's...um...not that simple...no extension, I think, but somehow the OS knows it's an executable. I don't really know precisely.
It's probably used here so much more than other places because the source code is (almost?) always available, and there is much variation in OSs and bits (e.g. Win32, Win64, Linux32, Linux64, etc. would all have the same source code but different binaries).

Yep, pretty much as Mini-Geek said.

[IronBits]

Basically, a non-text file vs. a TEXT based file.