[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 588.

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: spiccioli on February 26, 2013, 04:00:03 PM

Quote from: Aseras on February 26, 2013, 03:46:01 PM

Quote from: Subo1977 on February 26, 2013, 01:48:10 PM

Bo One here who Discuss about the l word?

It's pretty obvious something is wrong. normal variance should not be so cyclical.

http://p2pool.info/luck/

To be honest even this "something is broken" issue is very cyclical ... Wink

Very true. It's still the worst cycle I've ever seen. I said that before and organ proved it's been worse. I've wonder if there still have been worse days.

M

spiccioli

legendary

Activity: 1379

Merit: 1003

nec sine labore

Quote from: Aseras on February 26, 2013, 03:46:01 PM

Quote from: Subo1977 on February 26, 2013, 01:48:10 PM

Bo One here who Discuss about the l word?

It's pretty obvious something is wrong. normal variance should not be so cyclical.

http://p2pool.info/luck/

To be honest even this "something is broken" issue is very cyclical ... Wink

That said:

- memory consumption seems to happen less when using stratum (I've been using it for two days now and memory is stable around 400 Mb)
- upload traffic and "peers sent a transaction already known" should be addressed, if possible.

spiccioli

Aseras

hero member

Activity: 658

Merit: 500

Quote from: Subo1977 on February 26, 2013, 01:48:10 PM

Bo One here who Discuss about the l word?

It's pretty obvious something is wrong. normal variance should not be so cyclical.

http://p2pool.info/luck/

Subo1977

sr. member

Activity: 344

Merit: 250

Flixxo - Watch, Share, Earn!

Bo One here who Discuss about the l word?

rav3n_pl

legendary

Activity: 1361

Merit: 1003

Don`t panic! Organize!

Another lead to find memory eating routine: see my graph: http://rav3n.dtdns.net:9332/static/graphs.html?Day
From about 3PM my miner stuck and send invalid works. In same time memory usage was growing more and more and not get down when I restart miner and bad to mining.
Maybe there is something wrong on stratum/miner connections like it was in p2p connections?
My node eats now 420MB and still slowly rising.
After restart w/o shares after it re-download share chain it is about 100MB less.

Subo1977

sr. member

Activity: 344

Merit: 250

Flixxo - Watch, Share, Earn!

Quote from: Prattler on February 25, 2013, 04:49:48 PM

Quote from: Subo1977 on February 25, 2013, 04:10:25 PM

at my side its Ubuntu 12.10 X64 with all patches and python 2.7

Please run

Code:

python
import twisted
print twisted.version

and report the output.

>>> import twisted
>>> print twisted.version
[twisted, version 12.2.0]
>>>

Rogue Star

member

Activity: 89

Merit: 10

Quote from: Prattler on February 25, 2013, 04:49:48 PM

Please run

Code:

python
import twisted
print twisted.version

and report the output.

Sounds pretty much like my setup and problems, so it should be

Quote

>>> print twisted.version
[twisted, version 12.2.0]

Prattler

full member

Activity: 192

Merit: 100

Quote from: Subo1977 on February 25, 2013, 04:10:25 PM

at my side its Ubuntu 12.10 X64 with all patches and python 2.7

Please run

Code:

python
import twisted
print twisted.version

and report the output.

Subo1977

sr. member

Activity: 344

Merit: 250

Flixxo - Watch, Share, Earn!

Quote from: Prattler on February 25, 2013, 03:51:43 PM

Quote from: forrestv on February 25, 2013, 11:50:15 AM

Aseras/Subo1977: Are you using P2Pool 11.2 (the current newest version)? It fixed a memory leak that caused similar problems.

Perhaps issue a warning if twisted.version < 11.1? Because that could explain it...

at my side its Ubuntu 12.10 X64 with all patches and python 2.7

Prattler

full member

Activity: 192

Merit: 100

Quote from: forrestv on February 25, 2013, 11:50:15 AM

Aseras/Subo1977: Are you using P2Pool 11.2 (the current newest version)? It fixed a memory leak that caused similar problems.

Perhaps issue a warning if twisted.version < 11.1? Because that could explain it...

Aseras

hero member

Activity: 658

Merit: 500

Quote from: forrestv on February 25, 2013, 11:50:15 AM

Aseras/Subo1977: Are you using P2Pool 11.2 (the current newest version)? It fixed a memory leak that caused similar problems.

Yes 11.2 and .7.2

Subo1977

sr. member

Activity: 344

Merit: 250

Flixxo - Watch, Share, Earn!

Quote from: forrestv on February 25, 2013, 11:50:15 AM

Aseras/Subo1977: Are you using P2Pool 11.2 (the current newest version)? It fixed a memory leak that caused similar problems.

Yes actual p2pool and bitcoind from git

forrestv

hero member

Activity: 516

Merit: 643

Aseras/Subo1977: Are you using P2Pool 11.2 (the current newest version)? It fixed a memory leak that caused similar problems.

Subo1977

sr. member

Activity: 344

Merit: 250

Flixxo - Watch, Share, Earn!

Quote from: Aseras on February 25, 2013, 10:19:17 AM

There's definitely something going on in the background. After 5-7 days my latency goes way up and it slows everything down and I start getting much more dead/orphans. By 9 days p2pool goes almost all dead work. A reboot/restart fixes it all. Instantly. I'd suggest just putting in a switch or script that makes p2pool restart ever 24-36 hours as a temporary "fix/workaround"

+1

My Memory gets blown up... and my income goes down....

taltamir

full member

Activity: 196

Merit: 100

Quote from: Aseras on February 25, 2013, 10:20:26 AM

Quote from: taltamir on February 20, 2013, 07:59:58 PM

How do I make p2pool work for both bitcoin and litecoin simultanously?

change a port in the conf. both try to use 9332.

thanks

Aseras

hero member

Activity: 658

Merit: 500

Quote from: taltamir on February 20, 2013, 07:59:58 PM

How do I make p2pool work for both bitcoin and litecoin simultanously?

change a port in the conf. both try to use 9332.

Aseras

hero member

Activity: 658

Merit: 500

There's definitely something going on in the background. After 5-7 days my latency goes way up and it slows everything down and I start getting much more dead/orphans. By 9 days p2pool goes almost all dead work. A reboot/restart fixes it all. Instantly. I'd suggest just putting in a switch or script that makes p2pool restart ever 24-36 hours as a temporary "fix/workaround"

zvs

legendary

Activity: 1680

Merit: 1000

https://web.archive.org/web/*/nogleg.com

Quote from: Rogue Star on February 19, 2013, 08:34:58 PM

KC has asked me to post this on his behalf as he has yet to be white-listed. I haven't had a chance to investigate or even speculate if the same issue is affecting me, but perhaps this info will be useful to others in the meantime.

Quote from: KC_Atlanta

Hi Rouge Star,
I'm new to the forum and still in the newbies sandbox, so I can't post this, but wanted to share with you in hopes it helps. If it does, please repost to the forum as it may be a while before I can get the information up there myself. I'd spent some time looking to see if this had already been covered but found nothing related.

In a nutshell, I'd spent 4 days trying to figure out why my stale /orphan and rejected rates were abnormally high and had been troubleshooting as a network issue. My ISP forces me to do some extra work to set up NAT, so I'd double checked my NAT and FW rules verify that port 9333 was forwarding correctly and was sure it was.
From the machine I run P2Pool on, I fired up wireshark and filtered on tcp 9333 to make sure it was working asynchronously when I noticed a problem with numerous packets returning invalid IP and TCP checksums.
You may already know, most new NIC's by default will perform TCP, IP and UDP checksum calculations on board and I've run into issues with this especially with NAT where a firewall will drop the packet due to the invalid checksum. I disabled all of the "offload checksum" features on the NIC, and have had a huge improvement overnight with only 1 Stale, 2 orphans and <1% rejected in the last 12 hours.

Best of luck,
-KC

not because of any firewall, but I do this on all my systems as well (w/ netsh in windows and ethtool in linux)... unless I'm missing something here, the offloading is only beneficial if your CPU is overworked... otherwise it's faster to just not do any offloading (don't need to offload if your CPU is at 99% idle?). I disable RSS, chimney, and netdma in Windows as well... Leave DCA on unless I'm having CPU issues (then it goes off and some of the others come on, like NetDMA and offloading)

taltamir

full member

Activity: 196

Merit: 100

Thank you

rav3n_pl

legendary

Activity: 1361

Merit: 1003

Don`t panic! Organize!

Quote from: taltamir on February 20, 2013, 07:59:58 PM

How do I make p2pool work for both bitcoin and litecoin simultanously?

RTFM? It is described in readme....

Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 588. (Read 2591920 times)