Pages:
Author

Topic: [ Bonus PPS Pool - 130% PPS - OPEN ] - page 43. (Read 94030 times)

hero member
Activity: 504
Merit: 502
Cool, seems CGMINER 2.4 just got released, hopefully no surprises for me Wink
I definitely had services like yours in mind when I developed it, so I hope not   Wink

Great, good to hear I am somewhat on your development radar even though Im sure most of the pool issues is my own stupidity Smiley
-ck
legendary
Activity: 4088
Merit: 1631
Ruu \o/
Cool, seems CGMINER 2.4 just got released, hopefully no surprises for me Wink
I definitely had services like yours in mind when I developed it, so I hope not   Wink
hero member
Activity: 504
Merit: 502
Cool, seems CGMINER 2.4 just got released, hopefully no surprises for me Wink
hero member
Activity: 504
Merit: 502
UPDATE:

I had to stop the remote stats importing (its still logging, just not updating to webpage since its updating remotely) perhaps for the next hour or two, so far it seems to have fixed a huge socket leaking but need to verify that it indeed help resolve part of the issue.

As soon as I am sure that it resolved it, I will move the DB to the main pool to avoid network updates, expect an update in this thread  bit later about this. Once I update webstats you will have all shares accounted and updated, everyone can rest assured.

EDIT: Had to disable it again, still tracking some smb socket leaking, seems Im hitting every possible and probably unknown networking issue o_0
hero member
Activity: 504
Merit: 502
This might be pure coincidence with the efforts you are currently making to stabilize your pool, but with a cgminer build from latest GIT I have a noticeably better experience.

With the previous cgminer-2.3.6 my miner was falling back to the secondary pool most of the time during last few days. Plus I noticed that my router reported hundreds of active connections to the pools. With the latest updates Con added (reuse of existing connections, network scheduling improvements) that number came down to like a dozen and I'm back at your pool more reliably.

You'll need to pull the latest updates from today and build cgminer for yourself, or wait for 2.3.7.

I think its more related to things I changed my side and the blocking of leaking IP's Wink

But update could none the less improve things Im sure, just not the main reason right now Cheesy
donator
Activity: 919
Merit: 1000
This might be pure coincidence with the efforts you are currently making to stabilize your pool, but with a cgminer build from latest GIT I have a noticeably better experience.

With the previous cgminer-2.3.6 my miner was falling back to the secondary pool most of the time during last few days. Plus I noticed that my router reported hundreds of active connections to the pools. With the latest updates Con added (reuse of existing connections, network scheduling improvements) that number came down to like a dozen and I'm back at your pool more reliably.

You'll need to pull the latest updates from today and build cgminer for yourself, or wait for 2.3.7.
hero member
Activity: 504
Merit: 502
Bugger, had to readd some of the IP's, this is frustrating as hell.

Again the same IP's leaking, turning out to be a biatch to figure out.
legendary
Activity: 1204
Merit: 1000
฿itcoin: Currency of Resistance!
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.

My IP is on your blacklist... Anyway, do I still need to add this --no-submit-stale cgminer option?

Tks!
Thiago

Thats up to you, I just think it adds more stales than accepted shares so thats only reason I suggested disabling it.

Ive removed your IP just now from blacklist.

Online again!! Thank you!! I'll activate this option ASAP...

Thank you!
hero member
Activity: 504
Merit: 502
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.

My IP is on your blacklist... Anyway, do I still need to add this --no-submit-stale cgminer option?

Tks!
Thiago

Thats up to you, I just think it adds more stales than accepted shares so thats only reason I suggested disabling it.

Ive removed your IP just now from blacklist.
legendary
Activity: 1204
Merit: 1000
฿itcoin: Currency of Resistance!
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.

My IP is on your blacklist... Anyway, do I still need to add this --no-submit-stale cgminer option?

Tks!
Thiago
hero member
Activity: 504
Merit: 502
UPDATE:

So far Ive removed everyone except 3 IP's from the blacklist, still need to see what to do about them other than that connection been behaving. The only problem that is still there is some LP stale issues so Im sure most would still average out between 1-2% stales which I want to lower still but need to find the LP bug among other things.
hero member
Activity: 504
Merit: 502
I don't think they get rejected all the time, I just happened to see one pass by that was accepted.

 [2012-05-02 16:00:58] LONGPOLL from pool 0 requested work restart
 [2012-05-02 16:01:07] Accepted 162e43d6.74e5dd18 GPU 1 pool 2
 [2012-05-02 16:01:10] Accepted 77b073fd.64acfb4f GPU 0 pool 0
 [2012-05-02 16:01:10] Stale share from pool 0, submitting as user requested
 [2012-05-02 16:01:15] Accepted 9d6e1c35.ef3eacb7 GPU 1 pool 0

Im sure some get accepted, I just know more get rejected than accepted by far.
legendary
Activity: 1274
Merit: 1004
I don't think they get rejected all the time, I just happened to see one pass by that was accepted.

 [2012-05-02 16:00:58] LONGPOLL from pool 0 requested work restart
 [2012-05-02 16:01:07] Accepted 162e43d6.74e5dd18 GPU 1 pool 2
 [2012-05-02 16:01:10] Accepted 77b073fd.64acfb4f GPU 0 pool 0
 [2012-05-02 16:01:10] Stale share from pool 0, submitting as user requested
 [2012-05-02 16:01:15] Accepted 9d6e1c35.ef3eacb7 GPU 1 pool 0
legendary
Activity: 1260
Merit: 1000
That's not quite right...
With the new behavior of CGMiner/BFGMiner, when an LP hits on ANY of your pools, your CGMiner will start looking for new work.  If say Bonuspool is diametrically opposite from the pool that found the block (and that you happened to be connected to as well), the miner will find out about the LP before you will.  Therefore, a share might be stale taking the entire BTC network into account, but it can be valid on an individual given pool for several seconds to minutes depending on the connectivity of the pool. 

That's why the default behavior was changed and anyone the least bit concerned with efficiency is not going to turn of submission of stales, since it could, in some cases, significantly reduce your submitted shares.
hero member
Activity: 504
Merit: 502
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.
I think it's because stales are usually valid in P2Pool. (Obviously doesn't apply here)

Its still a fairly odd thing to enable by default if it applies to only p2pool thats why I am confused with the switch from off by default to on by default.

My understanding was that it had to do with the new LP behavior. If I'm mining at your pool but I detect the new block from Slush or Deepbit before I get it from you, the share is stale but your pool might still accept it.

Fair enough, but all I notice with this behaviour is rejected shares 100% of the time and many users are critical over stale percentages thus this doesnt help the visual appearance for having stales submitted after LP's only to cause gauranteed rejects.

That said, getting an LP from another pool still means bonuspool also received the LP and is on the new block allready, you are simply gambling that the active pool is slower to receive confirmation of new block which is a silly approach in my books, if there is a new block then dump old work and start with new work, thats the only approach I know of that works reliably.
legendary
Activity: 1274
Merit: 1004
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.
I think it's because stales are usually valid in P2Pool. (Obviously doesn't apply here)

Its still a fairly odd thing to enable by default if it applies to only p2pool thats why I am confused with the switch from off by default to on by default.

My understanding was that it had to do with the new LP behavior. If I'm mining at your pool but I detect the new block from Slush or Deepbit before I get it from you, the share is stale but your pool might still accept it.
hero member
Activity: 504
Merit: 502
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.
I think it's because stales are usually valid in P2Pool. (Obviously doesn't apply here)

Its still a fairly odd thing to enable by default if it applies to only p2pool thats why I am confused with the switch from off by default to on by default.
rjk
sr. member
Activity: 448
Merit: 250
1ngldh
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.
I think it's because stales are usually valid in P2Pool. (Obviously doesn't apply here)
hero member
Activity: 504
Merit: 502
UPDATE:

BTW, if you run cgminer please make sure you set --no-submit-stale on your miners, the default with the latest is to submit stales and it will obviously just return reject/stale for you, the proper handling is to dump stales, no idea what the reason is for enabling it by default in cgminer unless someone can enlighten me.
hero member
Activity: 504
Merit: 502
I'm out...  Sad

I'm using cgminer 2.3.6...

Let me know if there is anything I can do to help with this tests...

Thanks for your hard work!

Best,
Thiago

Heya, Im sorry but its just temporary, I am trying to get to bottom of this and dedicating alot of time atm to resolve it.

All I can figure atm looking at the users whos connections is leaking on the pool it seems to all relate to having alot of miners connecting over a single ip to the pool, its a strange problem but atleast its something to work from for me to fix it.

I have 6 rigs connect to BonusPool using the same IP... Maybe, you can try to make one worker account for each of my rigs to see if things behave differently...

Heya, its not worker related but seems to be gpu(active connections) coming in from one IP and that is what I am working on atm, ive allready let most connections back on with only 4 IPs still flagged and hopefully released tonight still, wont know until I work through it.
Pages:
Jump to: