Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 499. (Read 2591928 times)

I would check also if your HDD haven't got any bad sectors. Check your system logs and scan your HDD.

Check for cron jobs that fire at 15 past every hour, it could be that the server you are running it on is a bit underpowered in some way (most likely disk IO bandwidth) and it can't keep up.

My friend and I started a p2pool BTC node and every 65 minutes or so we get about 4 minutes of downtime.
See the graph here:
http://108.161.129.247:9332/static/

What could be causing these downspikes?

Thanks for the great idea/community everyone!

best is github version

why running apache public instead of p2pool? thats like having a more secure system (p2pool) and replace it with somethin vulnerable (apache2)...

Nice!

Is the "data" folder the only folder that should be copied from the old to the new, or are there more?  Does the p2pool program write to any files outside the "data" folder?  I should save a before-and-after copy so that I can compare and see what it changes.

Good news, it's settled down now, and starts to work.  It now sees the rest of the pool.  Thanks for letting me know that the timeouts/errors are normal when downloading the initial sharechain.

Josh

This might be the wrong place to ask, but I'm wondering if anybody else has done this.

I don't want my port 9332 open to the world (p2pool Web stats and RPC interface).

I have a Web server on my machine, so tried to redirect it, so that this would work:

URL typed into browser of https://mysite.example.com/whatever/p2pool/ --> proxy to http://127.0.0.1:9332/ on the inside

Although I wanted to keep port 9332 closed from the outside world, I still want to be able to peek at it remotely, and so have added https and HTTP basic authentication to it (old style username/password) on my Web server's configuration.

This works, but unfortunately it's fooled by the JavaScript and HTML in the page, having lots of links to places like "/static" and "/local_stats" and such.

I know my Apache proxy is working, because I get p2pool's redirections, such as when you go to port 9332 without a location and it redirects you to the "/static" webpage: https://mysite.example.com/static

This doesn't work on my setup, because I need to have p2pool be under a path, not at the toplevel.

So, I tried the "mod_proxy_html" Apache module.  This seems to be failing to parse the HTML generated by p2pool, and rewriting it correctly, because then, after doing that, I no longer get toplevel redirects, but I get p2pool to tell me "No such resource/No such child resource".  This text is an error from p2pool, not my Web server, so I know that at least the proxy is getting through to p2pool and making the internal connection to port 9332 correctly.  However, something's getting garbled in my attempt to rewrite the HTML, so it's not getting forwarded to the correct URL, I think.  Unfortunately my debugging trail goes cold there, I have no visibility into the internal workings of mod_proxy_html unfortunately.

Has anybody else confronted this problem?

I wonder if it would be possible to rewrite the links in p2pool's Web interface to be relative, not absolute, paths?  Would there be a technical problem with this, or would it be a good improvement to have?

Thanks!
Josh

When I upgrade, I usually unzip the new version into a new folder, move the data folder from the one to the new one, rename the old folder, then name the new folder the proper name (so my script still works).

As for the connections lost .. p2pool doesn't appear to be multithreaded.  When downloading the sharechain it gets bogged down in processing the shares that other things don't get processor power, so when it's done processing the shares, timeouts/errors pop up all over the place.

I always use the latest .zip from github.

M

Just upgraded p2pool from 13.1 to 13.2.

It's downloading the sharechain again now, and seemingly having a rough time of it: there's lots of Python exceptions all over the place, about connections lost.  My Internet connection is reasonably good, though, so unsure why it would be doing this (13.1 was fine).

What's the preferred way to upgrade from 13.1 to 13.2?  Is there a way to do it without losing the downloaded sharechain?  I just did a rough upgrade, unpacking a new directory and then running Python in there, as if I had never used p2pool before.

Also, what version is best to use?  The tarball release of 13.2 from github?  The master branch from github, cloned with "git clone"?
 
Thanks!
Josh

It wasn't just a flood of hash power, it was a bug in the difficulty adjustment someone was exploiting so they could mine blocks for their attack at 5K difficulty even if everyone else was on a far higher difficulty. Can't win that battle. Hash power coming and going isn't a problem for TRC because of the EMA difficulty adjustments, which were added earlier this year when hash power could mess things up.

Done, thanks!

rename thread, 2000/2500 GH/s

Currently on block 175055. Have you updated the latest build?

No, it is the result of having a pump and dump scam coin with little network support ...

TRC was flooded with hashpower, I think they 51%'ed it as well to orphan other miners blocks.  The difficulty was jacked up through through the roof, then then left.

https://bitcointalksearch.org/topic/terracoin-attack-over-12th-attack-confirmd-261986

This is the downside of short interval difficulty adjustment.  The two week one on bitcoin may be annoying but the short ones are open to far more abuse.

There is definitely something strange with Terracoin, I continue having troubles with TRC and P2pool but today the Terracoin wallet is unable to synchronize with the network from about 5 hours ago and it's stuck in block 175050.  

And I'm back to the Celeron machine.  I set CPU Affinity to bitcoin-qt to one CPU core and p2pool.exe to the other, and I'm getting sub-0.2s latency.  Job jobbed.

Yeah, I think there's something odd going on with Server 2012.  It's hovering about 0.7s at the moment.  Ran much better on Windows 7 on the Celeron 847. 

I think I'm going to abandon the idea of running the pool node on this machine.

Problem is on your side. Fix your system/software.
I have 0.2s on slower AMD APU machine (Linux AMD64).

I just tried running my p2pool node on my "main" server, which is a FX8320.  8-core 3.5GHz.  12GB RAM.  SSD driver.  Getwork latency is 2 seconds compared to 0.3-0.5 on my Celeron machine.  Good effort, AMD.