LLRnet and PRPnet servers for automated LLR

[Oddball]

The server seems to be down again

[mdettweiler]

Originally Posted by Oddball

The server seems to be down again

Hmm, it works for me. Additionally, I've been navigating the noprimeleftbehind.net web pages for the last 5 minutes or so before you posted this, and didn't notice any problems...strange.

Is it still down for you, or is it working again now?

[Oddball]

Originally Posted by mdettweiler

Hmm, it works for me. Additionally, I've been navigating the noprimeleftbehind.net web pages for the last 5 minutes or so before you posted this, and didn't notice any problems...strange.

Is it still down for you, or is it working again now?

It's still down for me, and I don't know why. Neither the LLRnet server or the PRPnet server works (the error code for the PRPnet server is 10060). The http://www.noprimeleftbehind.net/ website and the http://www.noprimeleftbehind.net/tps sites both don't work, but I can get to other sites (Yahoo, Google, Wikipedia, etc).

I'll restart the computer and delete browsing history and cookies to see whether that solves anything.

edit: Didn't need to do that, everything's working fine now. I wonder what happened earlier

[mdettweiler]

Originally Posted by Oddball

edit: Didn't need to do that, everything's working fine now. I wonder what happened earlier

Hmm, would you know. The best explanation I can come up with for this would be DNS troubles at your ISP; that can cause outages to some sites but not others. ISP DNS issues are usually resolved pretty quickly (compared to other types of outages), which would be consistent with what happened here.

[MooMoo2]

The PRPnet server has dried out.

I'll be back when more candidates from the most recent sieve file are loaded onto port 12000.

[mdettweiler]

Originally Posted by MooMoo2

The PRPnet server has dried out.

I'll be back when more candidates from the most recent sieve file are loaded onto port 12000.

Ah, thanks for the heads-up--I had been monitoring the server and was aware that it was going to dry out soon but it's good to know that it will be OK for you (currently the only active PRPnet contributor) to wait for a new sieve file to reload it. I was initially going to just load more from the current file, inefficient as that may be.

[Oddball]

The new small k file is now available:
http://www.sendspace.com/file/ul26bb

Max, could you load n=481501-482000 into the PRPnet server?

[mdettweiler]

Originally Posted by Oddball

The new small k file is now available:
http://www.sendspace.com/file/ul26bb

Max, could you load n=481501-482000 into the PRPnet server?

Done.

[cipher]

Just a thought. We should be tracked on NPLB's home page, with the point system. Also we should be tracked or get credit on
http://stats.free-dc.org/stats.php?page=proj&proj=nplb

Just a thought. I know twin prime is not an NPLB project how hard would it be to convert the existing points to NPLB and maybe being on NPLB home page we can attract new participants.

Cipher.

[mdettweiler]

Originally Posted by cipher

Just a thought. We should be tracked on NPLB's home page, with the point system. Also we should be tracked or get credit on
http://stats.free-dc.org/stats.php?page=proj&proj=nplb

Just a thought. I know twin prime is not an NPLB project how hard would it be to convert the existing points to NPLB and maybe being on NPLB home page we can attract new participants.

Cipher.

Actually, I have given some thought to this idea before (in a PM discussion with Oddball)...what I was thinking is that TPS could have their own distinct stats system based on the same underlying code as NPLB's. I've been meaning to talk to Dave, our stats admin, about this but haven't yet gotten the chance...been kind of busy in "real life". It's on my to-do list...

[mdettweiler]

Hi all,

Over in the LLR reservation thread, I just posted the reservation of the last n=480K-485, k<100K range, which has been loaded into PRPnet. This should last us until the end of the rally and probably won't take too long to polish off after that.

At that point, I will need to dry out the server so I can upgrade it to the latest PRPnet server version (4.1.3 as of this writing). The "normal" upgrade process going from 3.3.x to 4.0.x is rather arduous (I tried it on some personal servers and it was rather messy), so it will be much easier for me (and with less chance of error) to simply dry out the server, then start up a completely fresh one under version 4.0.x. Not to worry, though, all stats will be copied into the "new" database when I do this. Once I've done that, I'll upgrade the 4.0.x server to 4.1.x (a much simpler process that only consists of a couple quick SQL commands). The total downtime shouldn't be too long; the main thing for the end user will be that the server will be out of work for a day or two.

That done, I will load the next range (n=480K-485K, k=100K-200K) into PRPnet as discussed in the "What's next?" thread. As also discussed there, LLRnet will proceed on to n=485K-490K, k<100K. It will not be affected by the brief downtime on PRPnet.

Hopefully this will not cause any problems for anyone. On the plus side, 4.1.x will give us a number of cool new features:

-Built-in support for teams on the server stats page (should make things much simpler for Oddball in the next rally )
-Automatic testing of +1 counterparts for -1 primes (this is technically present already in 3.3.6, but it didn't work well for TPS since it tested +1 first instead of -1 and didn't save secondary residues)
-And a host of other features that don't really affect the way TPS will be using the server

Max