The stats server won't be able to run updates every hour like it does now without the copy-off/converter going every hour...

If we had the conversion script here, we could take the raw file, convert and import in one shot, every hour.

I know I asked for the log file format a few weeks ago and still waiting so we could do the conversion/import/hourly and be ahead of the game, but, looks like we are behind the 8 ball now.

Edit: saw Kar_bon's input, sounds like a much better solution to me - business as usual with the hourly copy-offs :smile:

[quote=kar_bon;162438]Max, your conversion is quite good but to much complicated for the purpose!

think about this:

because Rogue(Mark) wrote the PRPnet-client/server, i think he's able to include an option like

-L make LLRnet-server ouput-format (3 lines per k/n-pair)

so you won't need any script or new ghost-LLRnet-server to convert them.
with those there would be more chances of fails like ghost-server crashes or script won't started correctly!

my opinion to the use of a new program![/quote]
Hmm, I wasn't aware of that option. Was it introduced in the latest version or something?

Anyway, this is perfect. I'll look into using that option and see about getting the server set up sometime later today. :smile:

However, the -L option only would help for the results files part of the task. PRPnet still doesn't produce an LLR-like knpairs.txt file listing only the k/n pairs remaining in the server; instead, its prpserver.candidates file simply marks each number as "completed" when it's done, and leaves it in the file. So, thus it would be a lot harder to write a script to get the same information from a prpserver.candidates file natively that we usually get from knpairs.txt for the status reports.

That's the main reason why I was thinking of using the "helper server" setup--because that would allow us to pull our status data from knpairs.txt as always. I admit it is a little on the clunky side, but the only other option I can think of would be to completely re-write my status page script for PRPnet, which would be very, very difficult due to the type of format it uses.

Oh, wait--I just remembered something! Last I heard Rogue was working on a built-in status page generator for the PRPnet server (sort of like what ECMnet does, since PRPnet is based on ECMnet). A while back I sent him some test data from the CRUS G3000 server so he could test that function; I don't know if he's managed to finish it yet. I'll go and ask over in the PRPnet thread at CRUS. :smile:

I don't know why I didn't think of that before--that would solve all of our status page problems completely! :smile: That, combined with the -L option for generating LLRnet-style results files, should mean that we don't have any need for any converter scripts or helper servers whatsoever. :grin:

Max :smile:

[quote=mdettweiler;162467]Hmm, I wasn't aware of that option. Was it introduced in the latest version or something?

Anyway, this is perfect. I'll look into using that option and see about getting the server set up sometime later today. :smile:[/quote]
[quote=kar_bon;162438]because Rogue(Mark) wrote the PRPnet-client/server, i [B]think he's able to include[/B] an option like

-L make LLRnet-server ouput-format (3 lines per k/n-pair)[/quote]
i.e. it's not included yet. Karsten's just commenting on that Rogue could include it.
:doh!:

[quote=Mini-Geek;162474]i.e. it's not included yet. Karsten's just commenting on that Rogue could include it.
:doh!:[/quote]
Oh, duh! *bangs head* Whoops, my bad--I was skimming people's posts again. :redface:

No problem--the conversion script, in fact, is the least problematic of all these systems. All it has to do is just feed in the input and output files on the command line, and every time I've tested it it's run perfectly. It wouldn't be hard to set it up in the copy-off script.

The biggest issue is trying to get a status page generated. However, as I mentioned earlier, I think Rogue is currently working on a built-in solution for that, so that is probably just the easiest way to go.

Let's hold off on the PRPnet server for now until there is more time for a detailed technical analysis.

Go ahead and load the 6 k's for n=600K-700K into LLRnet port G8000. In the mean time, you can take your time coordinating with people on the PRPnet server and getting it set up the way we want. When the drive has completed to n=700K, then we can load n>700K into a PRPnet server assuming that it is ready at that point.


Gary

Whew! :grin:

k=325 is now back open for reservations for the enitre n-range.

[quote=gd_barnes;162480]Let's hold off on the PRPnet server for now until there is more time for a detailed technical analysis.

Go ahead and load the 6 k's for n=600K-700K into LLRnet port G8000. In the mean time, you can take your time coordinating with people on the PRPnet server and getting it set up the way we want. When the drive has completed to n=700K, then we can load n>700K into a PRPnet server assuming that it is ready at that point.


Gary[/quote]
Okay, that sounds good. I agree, holding off on the PRPnet server would be best for now--at least until Rogue releases a new version of PRPnet that includes the status-page code. :smile:

Running this mini-drive on G8000 should work fine since I've already got all the status page linkups in place. I'll get the server loaded with 600K-700K later today. :smile:

Hi all,

I am pleased to announce that LLRnet server G8000 is now loaded up and ready for action. :grin: (Just to reiterate exactly what it's loaded with: it's got the range of n=600K-700K for the 6-k k=300-400 drive.) This is now reflected on the [url]http://nplb-gb1.no-ip.org/llrnet/[/url] status page, and it should show up at [url]http://stats.ironbits.net[/url] as soon as results are submitted.

I'll get it added to the LLRnet servers thread shortly.

Max :smile:

P.S.: Oops, I just noticed that the [url]http://nplb-gb1.no-ip.org/llrnet/[/url] page still shows port 8000 as "for future use" instead of as "NPLB k=300-400 mini-drive". Oh well--that's only a cosmetic blooper, I'll fix it tomorrow. :smile:

[quote=mdettweiler;162544]Hi all,

I am pleased to announce that LLRnet server G8000 is now loaded up and ready for action. :grin: (Just to reiterate exactly what it's loaded with: it's got the range of n=600K-700K for the 6-k k=300-400 drive.) This is now reflected on the [URL]http://nplb-gb1.no-ip.org/llrnet/[/URL] status page, and it should show up at [URL]http://stats.ironbits.net[/URL] as soon as results are submitted.

I'll get it added to the LLRnet servers thread shortly.

Max :smile:

P.S.: Oops, I just noticed that the [URL]http://nplb-gb1.no-ip.org/llrnet/[/URL] page still shows port 8000 as "for future use" instead of as "NPLB k=300-400 mini-drive". Oh well--that's only a cosmetic blooper, I'll fix it tomorrow. :smile:[/quote]


Cool! Just to clarify, we will be searching k=315, 321, 327, 333, 339, & 345 from n=600K-950K with the mini-drive. I'll pull a couple of cores off of port 4000 to put on it to get us going while I'm on my trip.

Max, there is a dead link on the status page. It is the one for "Today's results processing log" at the bottom of the page. The link is [URL]http://nplb-gb1.no-ip.org/llrnet/todays_processing_log.txt[/URL] Can you fix that at the same time you fix the heading for port 8000? Thanks.


Gary

[quote=gd_barnes;162552]Max, there is a dead link on the status page. It is the one for "Today's results processing log" at the bottom of the page. The link is [URL]http://nplb-gb1.no-ip.org/llrnet/todays_processing_log.txt[/URL] Can you fix that at the same time you fix the heading for port 8000? Thanks.[/quote]
Okay. I think I'll just remove the link--it was meant originally to link to a logfile similar to the one for David's servers, but my scripts work somewhat differently and thus simply can't produce the same information to put down in a log file nearly as easily (for example, they screen for primes during the 15-minute status page updates, not at the end-of-day results file copy-off). And besides, even David's servers don't have a log like that posted anymore. :smile: