Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 508. (Read 2591928 times)

You have two different addresses with shares in the current share chain.

And you can only expect the 24 hour payment estimate to match the actual payment you get from a block when the expected time per block is 24 hours.

p2pool mining is beginning to look like a regular pool mining, something about it is definitely off...

For some reason I got two payouts on the last two blocks (one normalish sized and one super tiny)?



Also, the payout on my node looks way low:

Payout if a block were found NOW: 0.01026656 BTC to ***. Expected after mining for 24 hours: 0.134 BTC per block.

This is running at ~7.5Ghash for well over a day, so it should be closer to the 0.134 value. It seems like the "payout if block were found now" is actually the smaller of the two payments that I got for the last couple blocks? What's going on here??

My ulimit is 1024.... am I supposed to be having problems?

"fc19" is Fedora Core 19, that's really new, I'm unsure why they don't have a higher limit by default.  On a modern Linux system, with a lot going on, you should give yourself as many open file descriptors as you can.

I don't use Fedora, but the file should still be in the same place (I hope).  Edit the /etc/security/limits.conf file, put this at the bottom, don't forget the stars in the far left column:


Get rid of any other lines in the file that have "nofile" in them.  Unfortunately, it seems the maximum is 65536, the system simply won't let you set a higher limit (even though I wanted to).

Then, log out and log back in.  You should have a much higher limit now, verify with "ulimit -n" as you said.

Also, in your text you pasted, it says (in the big loud message) your Bitcoin daemon isn't running.  Something's seriously wrong on your system.  You should always have 3 programs running: bitcoind (or Bitcoin-Qt), p2pool, and bfgminer (or whatever miner you use).

Josh

Hi, during last couple days (probably ?! after upgrade to 13.1) I often see this :
2013-07-18 06:34:56.904438 Error when requesting noncached value:
2013-07-18 06:34:56.904486 > Traceback (most recent call last):
2013-07-18 06:34:56.904543 > Failure: twisted.internet.error.ConnectBindError: Couldn't bind: 24: Too many open files.
2013-07-18 06:34:56.904581
2013-07-18 06:34:56.904751
2013-07-18 06:34:56.904791 Error when requesting noncached value:
2013-07-18 06:34:56.904838 > Traceback (most recent call last):
2013-07-18 06:34:56.904894 > Failure: twisted.internet.error.ConnectBindError: Couldn't bind: 24: Too many open files.
2013-07-18 06:34:56.904934
2013-07-18 06:34:59.594579 > ########################################
2013-07-18 06:34:59.594676 > >>> Warning: LOST CONTACT WITH BITCOIND for 3.0 hours! Check that it isn't frozen or dead!
2013-07-18 06:34:59.594719 > ########################################
2013-07-18 06:34:59.594764 P2Pool: 17366 shares in chain (17371 verified/17371 total) Peers: 19 (13 incoming)
2013-07-18 06:34:59.594795  Local: 0H/s in last 10.0 minutes Local dead on arrival: Expected time to share:
2013-07-18 06:34:59.594824  Shares: 0 (0 orphan, 0 dead) Stale rate: Efficiency: Current payout: 0.0175 BTC
2013-07-18 06:34:59.594857  Pool: 1359GH/s Stale rate: 11.1% Expected time to block: 23.0 hours
2013-07-18 06:34:59.596119 > Error while calling getmininginfo:
2013-07-18 06:34:59.596182 > Traceback (most recent call last):
2013-07-18 06:34:59.596218 > Failure: twisted.internet.error.ConnectBindError: Couldn't bind: 24: Too many open files.
2013-07-18 06:34:59.596474 > Error getting work from bitcoind:
2013-07-18 06:34:59.596515 > Traceback (most recent call last):
2013-07-18 06:34:59.596550 > Failure: twisted.internet.error.ConnectBindError: Couldn't bind: 24: Too many open files.
====
python-twisted-core-11.1.0-2.fc17.x86_64
python-2.7.3-7.2.fc17.x86_64
====
I believe I saw it somewhere - might be problem with ulimit -n,but I think I increased it to 2048 and it did not help.
Restarting p2pool helps.
Any ideas how to debug the problem or what to do ?
Shall I upgrade fedora (python-twisted-core-12.2.0-2.fc19.x86_64) ?
Thanks a lot for your time !

how to set the share difficulty and the worker difficulty? like this : username/share diff+N ?

You can override the dust-preventing difficulty with the "/" option, which sets the desired share difficulty ("+" option sets desired pseudoshare difficulty). Sorry about not adequately describing the new features. Documenting them will happen.

3x longer shares = 3x higher difficulty = shares last 3x longer = approx. same payout per share per before, just find them less often and they stick around for longer.

The dust payout feature would raise difficulty on people, but I think that is more evident on LTC than BTC. I think the difficulty of dust in BTC is something like 1.5 or 2, but it is something reasonably larger on LTC (since LTC difficulty is commonly < 1)

I think it has to do with the new “dust payout” elimination feature. It raises the difficulty enough to minimize transaction fees when spending mined coins. It appears you may be able to reduce the share difficulty again by adding “+n” to the miner username, where n is the desired share difficulty. However, you may incur heavy transaction fees in the future when you go to spend the mined coins.

Maybe someone can confirm this for sure. I've been finding it difficult to find detailed p2pool documentation lately for some reason.

maqifrnswa is right though, with SPREAD=3 the "paid 3x as many times" is only when p2pool has relatively low hashrate, we get 3 blocks in less than 3 days now.

With the current hashrate you should expect your shares to be paid for 3 times each.

You are paid for each share every time a block is found as long as the share is still "active".  When the pool is lucky, you'll get paid for your share many times.  When the pool is unlucky, you'll get paid fewer times or not at all if the pool is very unlucky.  Because your shares stay active for approx 3x as long now, you'll get paid for your share 3x as many times (on average).

So that means I am paid three times für one share?

you're right, it is 3 days EXCEPT spread is 3. so if time to block is 20 hours, shares only will last for 60

so
3x longer shares = 3x higher difficulty = shares last 3x longer = same payout per share per before, just find them less often and they stick around for longer

EDIT: what chaged - shares are 3x harder to find and live for 3x longer, although spread is 3 - and time to block is like 20 hours, so shares last 60 hours (instead of 72)

Odd, if it IS 3 days on 30sec longpool share diff should be same like in 1 day/10 sec...
What changed?

The chain is the same number of shares as before, but because shares are found 3x less often, that means that the chain is 3 times longer in real-word time (3 days instead of 1 day).

Like it or not, the bitcoin mining community is in the midst of a transition to ASICs and the result is that small miners (GPU miners) are going to be left behind, no matter what pool you are on.

Ahh, this is BAD, I was sure that chain lengh will be change too, that payout last long for up to 3 days. But it is NOT.
Now small miners are killed for p2pool.

After the cutover, the targets for finding shares is now only one per 30 seconds instead of one per 10 seconds.  That is accomplished by making share difficulty is 3x higher.


To get shares more often, the only thing you can do is add more hashrate.  If you can't or don't want to add more hashrate, then the only thing you can do is be content with the increased variance.  In the long run, you'll earn the same as before, it will just be more spikey (you'll find a share 3x less often, but each share will be worth 3x as much).

why is the diff so high in version 13.1?

9k is a little bit high isn't it?

I don't get any shares. What can I do?

the whole v11 nodes connecting then me requesting parent shares from like 10 nodes every few seconds is causing some issues i think

so i set it to 4 incoming,  usually 2 or 3 connect right away, hopefully last one or so won't be v11

i put a log file of the action up temporarily at www.nogleg.com/swank.txt

i think I got all the IPs,  guess it's not a huge deal