After some testing, with the help from denisx in the irc, we figured out that this issue seems to
bo related to pushpool, phoenix and perhaps nvidia, all combined.
He runs a pool too and when I've pointed my phoenixminer at his pool, the same results.
Now, when I do the same thing with poclbm against my and his pool I have no unknown-work, in fact ~0 stales.
Tested with different versions of phoenix and pushpool.
diablo-miner and poclbm are working great.
additionally denisx reported that some members of his pool with nvidia cards have problems too.
Topic: [ERROR] Pushpoold "unkown-work" $50 Bounty (Read 4043 times)
the problem is still present.
i pledge 50$ for a solution. I'd pay via paypal since I have not much BTC.
i pledge 50$ for a solution. I'd pay via paypal since I have not much BTC.
I've tried increasing the askrate and it helps a litte bit.
their are still 50% unknown-works, how can i check that pushpool logs everything right?
their are still 50% unknown-works, how can i check that pushpool logs everything right?
Is this the "auth.cred_cache.expire" : 75, option?
No.
Quote
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
somewhere else on the host?
The problem is with miners submitting old work. They should not submit work older than 60 seconds, much less 120 seconds.
increasing the askrate of the miner would help against this?
but why do I have no troubles on other pools?
Is this the "auth.cred_cache.expire" : 75, option?
No.
Quote
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
somewhere else on the host?
The problem is with miners submitting old work. They should not submit work older than 60 seconds, much less 120 seconds.
you're right.
but the problem is very strange.
i figured out it has to be something wrong with the log.
i've received work and submitted within 120seconds.
where does pushpool log this, in memcached?
but the problem is very strange.
i figured out it has to be something wrong with the log.
i've received work and submitted within 120seconds.
where does pushpool log this, in memcached?
Is this the "auth.cred_cache.expire" : 75, option?
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
I dont think so... i think this option is how long it caches the authentication credentials (login/pass) for the worker but could be horribly wrong
Havent looked at that part of the code yet. check config.c in the src and that should give you a clue. maybe also check that your system time is correct and you not drifting crazily? i dunno... im outta ideas for ya man.... that option i mentioned solved my issue (started when someone joined the pool with 30 GH/s)
Is this the "auth.cred_cache.expire" : 75, option?
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
And if not how do I get pushpool do keep it longer, or is the problem
somewhere else on the host?
i've played a bit aroung with pushpoold and everything worked just fine.
But when I asked a friend for a little test something strange happened.
I have about 100MH/s my friend has 1500MH/s.
As soon as my friend joined for testing i got a lot of stales,
I looked it up in the datase and the reason is "unknown-work".
But when I asked a friend for a little test something strange happened.
I have about 100MH/s my friend has 1500MH/s.
As soon as my friend joined for testing i got a lot of stales,
I looked it up in the datase and the reason is "unknown-work".
pushpool keeps a log of the last 120 seconds' worth of work received from bitcoind.
unknown-work error occurs when a miner submits work that is not in that log.
here is the config:
and thx for your help
Code:
{
"listen" : [
{ "port" : 8342 },
{ "port" : 8332, "protocol" : "http-json" },
{ "port" : 8335, "protocol" : "http-json",
"proxy" : "XXX.XXX.XXX.XXX" },
{ "host" : "127.0.0.1", "port" : 8338, "protocol" : "binary" }
],
"database" : {
"engine" : "mysql",
"host" : "localhost",
"port" : 3306,
"name" : "XXXXXXXX",
"username" : "XXXXXXXXXXXXXX",
"password" : "XXXXXXXXXXXXXXXXXXXx",
"sharelog" : true,
"stmt.pwdb":"SELECT password FROM pool_worker WHERE username = ?",
"stmt.sharelog":"INSERT INTO shares (rem_host, username, our_result, upstream_result, reason, solution) VALUES (?, ?, ?, ?, ?, ?)"
},
"memcached" : {
"servers" : [
{ "host" : "127.0.0.1", "port" : 11211 }
]
},
"pid" : "/home/XXXXXX/bitcoin/pushpool.pid",
"forcehost" : "localhost.localdomain",
"log.requests" : "/home/XXXXXX/bitcoin/request.log",
"log.shares" : "/home/XXXXXX/bitcoin/shares.log",
"longpoll.disable" : false,
"auth.cred_cache.expire" : 75,
"rpc.url" : "http://127.0.0.1:10000/",
"rpc.user" : "XXXXXXXXXXXXXXXXXXXX",
"rpc.pass" : "XXXXXXXXXXXXXX",
"rpc.target.rewrite" : true,
"roll.ntime.disable" : true
}
and thx for your help
yes I did, this is frustrating. I have no clue where this comes from.
can you post your server.josn file (edit the passwords and usernames out tho)
yes I did, this is frustrating. I have no clue where this comes from.
Thanks for the hint .
I'm now testing, with and without the patch suggested by redditorrex.
so far it's not looking that good :/.
edit: it got a little bit better.
.I'm now testing, with and without the patch suggested by redditorrex.
so far it's not looking that good :/.
edit: it got a little bit better.
did you upgrade to newest pushpoold 0.5.1 and newest bitcoind and restart them both after editing the server.json file ?
Thanks for the hint .
I'm now testing, with and without the patch suggested by redditorrex.
so far it's not looking that good :/.
edit: it got a little bit better.
.I'm now testing, with and without the patch suggested by redditorrex.
so far it's not looking that good :/.
edit: it got a little bit better.
Try adding this to your server.json config
Code:
# disable X-Rollntime support
"roll.ntime.disable" : true
Dont forget to add an extra comma at the end of the config option above it if you put this at the end of your config.
Hello,
i've played a bit aroung with pushpoold and everything worked just fine.
But when I asked a friend for a little test something strange happened.
I have about 100MH/s my friend has 1500MH/s.
As soon as my friend joined for testing i got a lot of stales,
I looked it up in the datase and the reason is "unknown-work".
Where does this come from?
Help with this would be really nice.
regards, redshark1802
ps: things I've tried so far with help from the irc-channnel
-checking that the blockchain is complete
-checking if the time is correct
-decreased the speed from my friend to 300MH/s
-my friend has no problems at all
-unknown is like 90% of the work sent
i've played a bit aroung with pushpoold and everything worked just fine.
But when I asked a friend for a little test something strange happened.
I have about 100MH/s my friend has 1500MH/s.
As soon as my friend joined for testing i got a lot of stales,
I looked it up in the datase and the reason is "unknown-work".
Where does this come from?
Help with this would be really nice.
regards, redshark1802
ps: things I've tried so far with help from the irc-channnel
-checking that the blockchain is complete
-checking if the time is correct
-decreased the speed from my friend to 300MH/s
-my friend has no problems at all
-unknown is like 90% of the work sent
Jump to: