Vladimir
|
|
June 25, 2011, 08:51:34 AM |
|
I keep getting "configure: error: Missing required libevent" in bsd. It's installed though.
You need full libevent devel package setup, not just the library. Jeff some more help on this one. I have no idea what you mean by "You need full libevent devel package setup, not just the library." Please explain.
|
-
|
|
|
Xenland
Legendary
Offline
Activity: 980
Merit: 1003
I'm not just any shaman, I'm a Sha256man
|
|
June 26, 2011, 08:48:12 AM |
|
I keep getting "configure: error: Missing required libevent" in bsd. It's installed though.
You need full libevent devel package setup, not just the library. Jeff some more help on this one. I have no idea what you mean by "You need full libevent devel package setup, not just the library." Please explain. I think he means you need the development package as well as the binary package
|
|
|
|
Vladimir
|
|
June 27, 2011, 02:37:29 PM |
|
I've been running pushpoold on my box for sometime and about every 12 hours of running or so it begins to error the error log says WARN too many open files anyone else run into any similar issues?
Yep same issue. On my volumes it lives for not more than 5 minutes than this. My software detects this within 2 seconds, cleans it up, kill/restarts pushpool. ulimit -n is 32k. Not helping much... The best I got with it is 5% of stales which is not acceptable. Back to older ways it is than, for now at least. Will see if I can find time to reimplement it with Erlang.
|
-
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
June 27, 2011, 02:45:26 PM |
|
blkmond is what cleans up the open files (actually connections).
It's not a great solution, but it's the way pushpoold was designed.
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
Vladimir
|
|
June 27, 2011, 02:51:16 PM Last edit: June 28, 2011, 07:30:13 AM by Vladimir |
|
blkmond is what cleans up the open files (actually connections).
It's not a great solution, but it's the way pushpoold was designed.
blkmod just sends USR1, is that correct? Or does it do something else? I do not use it, but I ahve some other software sending USR1 to pushpoold when a block is solved. Have I messed up something badly here? actually I used this (thought that a handful of lines in perl is better than a handful of pages of python): #!/usr/bin/perl
use common::sense ; use File::Tail ;
my $bitcoinLog = File::Tail->new("/s/bitcoind/db/debug.log");
while ( defined ( my $line = $bitcoinLog->read ) ) { $line =~ /ACCEPT/ or next ; foreach( split ( /\n/ , `ps aux | grep pushpoold | grep -v grep` ) ) { /^.*?(\d+)/ and $1 and say "sent SIGUSR1 to process $1" and `kill -USR1 $1` ; } }
It took me much less time to write the above code than even skim over blkmond's code.
|
-
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
June 27, 2011, 03:04:30 PM |
|
As far as I know, I think that is all it does. I haven't looked too closely at it. So you are doing that already and still having the files issue?
Have you changed your hard and soft file limits?
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
Vladimir
|
|
June 27, 2011, 03:12:44 PM Last edit: June 27, 2011, 03:23:43 PM by Vladimir |
|
As far as I know, I think that is all it does. I haven't looked too closely at it. So you are doing that already and still having the files issue?
Have you changed your hard and soft file limits?
/etc/security/limits.conf: root soft nofile 34096 root hard nofile 34096
Once I point lots of Ghps to this it does not live much longer than 5 minutes, perhaps simply not enough time between solved blocks.
|
-
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
June 27, 2011, 03:32:20 PM |
|
How much is a lot? It should be able to handle ~450 GH/s before buckling.
I have my notes at a home on things that need to be changed in addition to hard and soft file limits. I'm trying to recall them from memory but I'm drawing a blank.
Oh, it's also very dependant on the speed of the backend DB. If your backend db server is slow or over a slow link, it will start getting backed up and eventually never catch up.
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
Vladimir
|
|
June 27, 2011, 03:41:32 PM |
|
Oh, it's also very dependant on the speed of the backend DB. If your backend db server is slow or over a slow link, it will start getting backed up and eventually never catch up.
Good idea, will change the code to cut db backend balls out. I am happy with just using log files.
|
-
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
June 27, 2011, 03:46:46 PM |
|
I rewrote pretty much all of the DB code in my implementation of PP. The stock implementation is atrocious and unusable/unmaintainable.
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
jgarzik
Legendary
Offline
Activity: 1596
Merit: 1099
|
|
June 27, 2011, 06:03:39 PM |
|
Version 0.5.1 released.
This is a bug fix release, with two minor bug fixes:
- fix stderr logging - better configure script host detection
|
Jeff Garzik, Bloq CEO, former bitcoin core dev team; opinions are my own. Visit bloq.com / metronome.io Donations / tip jar: 1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj
|
|
|
phorensic
|
|
June 28, 2011, 04:09:25 AM |
|
All the new ideas and coding on the github site in the last two weeks and that's all we get in the official fork? Gahhh!!! ETA for v0.6? "When it's ready" amirite?
|
|
|
|
jgarzik
Legendary
Offline
Activity: 1596
Merit: 1099
|
|
June 28, 2011, 04:31:21 AM |
|
All the new ideas and coding on the github site in the last two weeks and that's all we get in the official fork? Gahhh!!! ETA for v0.6? "When it's ready" amirite?
The pull queue is basically empty, so I'm not sure what you're talking about.
|
Jeff Garzik, Bloq CEO, former bitcoin core dev team; opinions are my own. Visit bloq.com / metronome.io Donations / tip jar: 1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj
|
|
|
DaMan
|
|
June 28, 2011, 08:44:12 AM |
|
is there a tutorial for Centos 5 out there? I can't get it to work.
|
|
|
|
asper
Newbie
Offline
Activity: 22
Merit: 0
|
|
June 28, 2011, 01:24:55 PM |
|
I keep getting "configure: error: Missing required libevent" in bsd. It's installed though.
You need full libevent devel package setup, not just the library. Jeff some more help on this one. I have no idea what you mean by "You need full libevent devel package setup, not just the library." Please explain. Yes, the same problem - i've installed libevent (cd /usr/ports/devel/libevent && make install clean), but there's no use for pushpool to compile, the configure script stops, saying: checking for event_base_new in -levent... no configure: error: Missing required libevent Can somebody tell me, where can I get this "full libevent devel package setup" ?
|
|
|
|
Vladimir
|
|
June 28, 2011, 04:07:36 PM |
|
you could try to compile pushpool on 32bit linux and than run it in binary linux emulation mode on freebsd. One option is to use statify on it to avoid messing with linux libs on freebsd. I have not tried yet since all my linuxes are 64bit and just did not have time to get this sorted. Theoretically it might just work. Let me know how it goes if you try it.
|
-
|
|
|
jgarzik
Legendary
Offline
Activity: 1596
Merit: 1099
|
|
June 28, 2011, 05:41:17 PM |
|
Jeff some more help on this one. I have no idea what you mean by "You need full libevent devel package setup, not just the library." Please explain.
Yes, the same problem - i've installed libevent (cd /usr/ports/devel/libevent && make install clean), but there's no use for pushpool to compile, the configure script stops, saying: checking for event_base_new in -levent... no configure: error: Missing required libevent Can somebody tell me, where can I get this "full libevent devel package setup" ? Sounds like user error. INSTALL file covers the basics of how to use a standard GNU autotools configure script. A script cannot magically detect site-specific directories for software installation. If you have a bunch of stuff installed into non-standard directory /usr/local -- common on FreeBSD -- then you want something like CFLAGS="-O2 -Wall -g -I/usr/local/include" LDFLAGS="-L/usr/local/lib" ./configure
|
Jeff Garzik, Bloq CEO, former bitcoin core dev team; opinions are my own. Visit bloq.com / metronome.io Donations / tip jar: 1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj
|
|
|
Vladimir
|
|
June 28, 2011, 05:54:04 PM Last edit: June 28, 2011, 06:17:52 PM by Vladimir |
|
not everyone is c developer and familiar with autotools and it is not "site specific" it is OS specific BTW
|
-
|
|
|
martok
|
|
June 29, 2011, 02:22:40 AM |
|
Yep same issue. On my volumes it lives for not more than 5 minutes than this. My software detects this within 2 seconds, cleans it up, kill/restarts pushpool. ulimit -n is 32k. Not helping much...
The best I got with it is 5% of stales which is not acceptable. Back to older ways it is than, for now at least. Will see if I can find time to reimplement it with Erlang.
You're getting the stales because you're killing pushpoold. Pp keeps a log of what work it has issued and when a miner submits work to a newly restarted pushpoold, it won't accept the share because it is not in the pp log.
|
|
|
|
martok
|
|
June 29, 2011, 02:24:04 AM |
|
Have you changed your hard and soft file limits?
/etc/security/limits.conf: root soft nofile 34096 root hard nofile 34096
Once I point lots of Ghps to this it does not live much longer than 5 minutes, perhaps simply not enough time between solved blocks. Well that changes the limit for root. But you are not running pushpool as root right?
|
|
|
|
|