Yea, 14 is the most connections I can get with Windows-7 HOME, using the 8-connects + the 8-addnodes trick... Guess another socket or thread is needed to extend that range for us.

I think two connections are the loopbacks... so it actually is 16 total, 8 and 8. But the loopback consumes 2 connections for RPC I guess.

If you mined coins on a fork... they don't actually exist... the fork rewards were not correct. They will show as a "?" in your wallet. However, tx's that were sent, will just be removed, unless the TX exists on both chains, or the sender resends them. (If you never get them, they return to the sender, by expiring in the senders wallet.)

If you go to your HELP menu, and go to the DEBUG WINDOW, then select the CONSOLE tab... Type the following...

repairwallet

If you mined coins on the wrong fork, it will show... Repaired 40 blocks, 400 coins removed/fixed. Or something like that.

However, if you are still not on the correct chain... and mining... you will have to do that again, after you get back on the correct chain.

Like I said... Magic... (PRO), the unrestricted version of Windows.
- Ability to add languages
- Ability to have more RAM
- Ability to have more than 2 networks
- Ability to do remote-access
- Ability to have more than 8 persistent connections per socket/thread
- Ability to use GB netspeeds
- Ability to use backup and restore correctly
- Ability to have multi TB drives
- Ability to have multi-boot OS's
- Ability to use full (real) compression for video and audio
- Ability to see "other" updates from windows update

That, and tons more things that they purposely remove/restrict/limit/impede for home users. Oh, they have the ability... for a price... Only $300 more. and hours with tech-support to find out why it doesn't do this shit that the other versions do, and why we can't just turn it on, or download things to make them work... Oh, sorry Mr... It is because we purposely made that version so it stopped those things from functioning. What did you think we spent all our time here doing... degrading our OS for cheap bastards like yourself, who actually want what you paid for, each year we force you to upgrade by removing things that our OS once did, but now we want you to upgrade, and you didn't...

P.S. Seems to choke every 20,000 blocks, when downloading from 0 still... I knew I shouldn't have deleted that backup until I had a full connection to the new chain... WHY ME! lol.

P.P.S. With only 20,000 blocks left to download... it is choking every 1,000 blocks or less... Just hangs forever... With 14 connections. (Not sure still if it is them stalling on delivery, or if my wallet is just "confused" attempting to rebuild the database from scratch. It is hitting all my mined blocks, going crazy trying to update the wallet while the blocks arrive. Audit issues? From all the forks and "invalid blocks", "orphaned blocks", that are just hanging around in my wallet.)

Hush... you with your special magic and connection wizards!

This on your windows QT wallet? (Forgot, I asked you that already, you had PRO or ULTIMATE or something...)

YAY, I got 14 now... 8 addnodes and 6 connects!

Well, now I know it is possible to physically have more than 8... So there is something there that needs some work. Seems to allow some more than 8 addnode connections, and others it limits... That seems to defeat the limit, so it is not a port-forward thing. It is a code thing. Microsoft limiting home users to only 8 connects per thread/socket? (Call it a self-limiting thing that MS likes to do to everyone. This is not the only thing they purposely degrade, to make you pay more for the "pro" versions. They degrade everything now.)

Still hangs every few thousand blocks... like they get tired of sending blocks, or the program just hangs, not sure how to build until I restart and it continues where it left-off.

P.S. Seems that the wallet does "rescan" every time it loads... even if you don't have -rescan in startup or bat-launchers... makes restarting a pain in the ass... Please remove that from hard-coding, unless you also hard-code repairwallet too, it is useless to have it do at every startup.

Three forks now...

83074, 83093, and 83098

Just noticed something... lol...

I put max connections 8... and had 1 "connect"... Now I have 9 connections. (Most I ever get is 8, even when I tell it to connect to 30. Now I know how everyone is getting more than 8. They have x# of connects plus, the 8 addnodes.)

Thus, max connections should be 0 if you only want to use connect to connect to 1 node. Apparently, connections are separate to addconnections, and they are not counted towards the actual connections.

Again, I am seeing the correct height, but the database doesn't seem to be doing anything to "fix" itself. It just sits there telling me I need 18... 19... 20... 21... more blocks, as the blocks build, but no-one is sending them. Or they are sending them, but the wallet isn't doing anything with them... like repairing the database with the new correct-height chain data coming in.

(Or if it is, it isn't updating the information on the QT client, to indicate "synced" or updated, or anything.)

UPDATE: It is getting the blocks from the wrong chain to update... Yet, displaying the correct chain height and information...

EG, Chain with height 83069 is the blocks I am getting delivered, but it is using the height of the chain with 83075... Thus, never going to catch-up, as it keeps sending me blocks that don't belong to the higher chain. Which it would later have to remove, to rebuild the correct blocks for the taller chain it thinks it is getting them from.

EDIT: Damn-it... It stopped updating once it got the last block of the wrong chain... and I just deleted my backup chain, thinking this was the new chain coming in.

Mini-fork just started...

I was 8 blocks behind and connections dropped down to 4, of 8 for me...

Reloading to sync-up with block-crawler height.

EDIT: Also... Update on 3.3.1 CGminer... I am completely back to my expected 10% losses, actually down to about 5% losses from rejections at the moment. Must be in a good-luck streak. (Only applies to solo miners who submit only blocks, not pool miners who submit shares and blocks.)

Block-crawler is loosing connections fast, down to 5, but it is still on the correct chain, I believe.

I have that same chain with 8 solid connections. Still can't get more than 8 though. Height 82440 @ 9:57 AM (GMT -5)

Still seems as if POS isn't the issue. (Seems like it is one of the indicators of the issue.)

Are you sure there isn't a value like X>=Y but it should just be X>Y due to something like this...
Diff/65536, but the results are off by 1, due to bytes starting at 0 not 1 resulting in 0-65535, where something like the diff had a value added to it... like 1 to make it not-0 for a division formula that would normally failed with an error of division-by-zero.

Eg, we are submitting an honestly valid submission, that the program sees as invalid. (Eg, not meeting the required criteria of being equal-to-or-greater-than. Because every time it is actually equal, it appears to be invalid, because the result is being compared by the modified value which was not remodified to conform to the premodified value.

(Not saying it is that specifically. But as rare as it is to get an exact diff... That would make it appear to be sporadic and without any apparent rhyme or reason. Leading to what the program believes is an honest "misbehaving" client, leading to the mass genocide of connections.)

This would/may be hidden in things like TIME, nonce, hash-diffs... of the most obvious locations, but in an obscure formula... one that might simply shift-bits or may have a signed value passing into an unsigned value, or mismatched value that isn't acting as expected... That, or an issue with 32/64 bit formula variations failing, between OS's. Where the 64-bit formula works as expected, but it provides a different output on the 32-bit version of the formula, operating on a 64-bit system.

Again, I am just throwing things out there... Things that have often caught me in programming, that are not always obvious to see, unless you compare the data at every step. (Compare it as the program sees it, not as an "output", which is converted before displaying.)

Unless everyone turned on POS, on the block-crawler... and the non-POS miners all left "of free will"... Then it is safe to say that POS is not the ONLY thing causing the issue with connections. Though it did aid it to happen faster. POW blocks are higher diff, and rarely hit EQUAL expected values for submission. Less when the blocks are constantly being reset by POS submissions, and even less with multiple high diffs. (Though many POS blocks ARE the exact equal expected diff, because the diff is so low, and there are many more diffs of those lower values. Same will happen at the other end of the spectrum, when diffs are real high, but being further apart, the rejections will be fewer and more catastrophic to those mining, as they will lose more, because those rejected blocks will be high value.)

Ok, so CGminer version 3.3.1 seems to be the last one that functions correctly for scrypt. Apparently the programmer has no interest in scrypt coins, or offering any resolution for them. BTC ASIC lover... You will literally get a cold shoulder if you mention scrypt in that thread. Like solid ice-cold, cold.

Since the diff has gone up, above 0.70, and with the older 3.3.1 CGminer, I am back UP to getting 2-4 blocks per hour off of 3.2MHs. Which was down by 75% when diff was below 0.65, I was only getting about 1 an hour, if that...

I will chalk-it up to a combo of the two... low diff collisions and crappy CGminer version, submitting crap too late.

These versions of CGminer all seem to have the same issues (for scrypt), or worse...
3.3.2 (displays wrong info in accepted and rejected columns, trouble submitting solutions.)
3.3.3 (displays wrong info in accepted and rejected columns, trouble submitting solutions, trouble getting work.)
3.3.4 (displays wrong info in accepted and rejected columns, trouble submitting solutions, trouble with fan control.)

Also note... I had lower and lower hash-rates with each version step-up. (I switched to 3.3.1 originally from 3.1.0, because it gave me better hash-rates and control for my 7970's. Anything above 3.3.1 seems to only focus on ASIC things and FPGA things, degrading scrypt-mining in the process.)

I am almost up to 20,000 coins now... Thanks to the work that has been done so far. Thank-you again, and again, and again! (10,000 were purchased, not mined.)

3.3.2 version of CGminer seems to have the same issue of displaying the wrong info in the accepted and rejected columns... going back down to 3.3.1... lol...

Yes, I noticed that too... 1.4.1 and 1.4.2 often have to be manually terminated from task-manager. Seems to hang if I don't shut-down my miners first, and then shut-down the wallet. I assume the network components are still talking to the miners, causing the hang.


This is more of a "security" and "value saver"... thing... having a minimum diff.

With a minimum diff (eg, system rejects anything faster, including attacks of natural and forced nature, made easy by letting diff fall to near-zero, while mining alone.) Having a minimum diff forces an attacker to have tons of individual power, if a coin ever falls below that point. Having a minimum diff also stops massive dumps that degrade coin value, causing people to stay away and never join it again. Dumps by those lucky three or 20 people getting tons and they just keep dumping them for less and less, until value for those who hold would take about a year to obtain the losses.

I am not saying it is a perfect solution... Just a realistic economic one that takes operation of the method into consideration. Plus, it allows coins to distribute fast enough to limit super-stacks of isolated forks. The network would still function, because of POS, which would dominate or sustain at that lower diff, if miners decided to stop mining at that point. Which they would not. Someone would mine it. Hell, there are people still mining sexcoin and it doesn't even have an exchange! (I still do, with some of my miners. Gotta have sexcoins!)

Reverted back to cgminer 3.3.2, the last known "working" version that didn't cause issues with scryptcoins.

Will be nice to actually see the accepted and rejected qty again, instead of having to wait a day to guess if the network was screwing me or not. (I usually leave if diff drops below 0.65 on most coins. Because estimates of value become useless and they never display correct losses. Most losses are never reported or recorded, just thrown-out. Guess it never crossed their minds to actually try mining, and simply compare results with expected calculations, and see the reality of what you get. But then again, they are not in the business of making us money, only themselves. They like when we mine junk-coins that don't actually reward. Because they mine the ones that actually do. The #2, #3 and #4 coins on the lists. xD.)

Anyone bored and want to walk me through compiling one of these coins on windows, from source? Never used QT before, and not sure what they are using to compile them in the first place. But I am a fast learner.

Hmm... maybe it is this new cgminer that is causing the issues for me then... (Unless I am just real far from the source and you are closer to the ones generating POW and POS.)

My cgminer shows the diffs of the rejected and accepted blocks, not the quantity of accepted and rejected. (Bug in 3.3.4) I based my results off actual results, the end of the day yields.

At the half-diff, I am getting 50% less "accepted" in cgminer, and 75% more "rejected", as opposed to simply having 50% more "accepted" and 50% more "rejected". I am getting more solutions, but 75% of them are lost to another faster block.

I will fire-up the older cgminer, since this one seems borked. If it isn't displaying the correct values, I assume more inside may be borked too.

But the statement about having min diff for POW is still a strong conviction with me. It hurts no-one to do that, and only helps to ensure coin value stays high. (Even in the event of 90% of everyone leaving, where 3 people end-up getting all 100% of the coin, now they get the same amount as if there were hundreds of people. But if the coin came to that, it wouldn't matter at that point.) This is from experience mining fast coins, like fastcoins, sexcoins, mincoins, luckcoins, memecoins, anonymouscoins, etc... You can look at the charts and see the issues. You actually get more mining with higher diffs than with lower diffs, due to the major losses. On average. Unless you are uber-lucky and one of the sources of, "we just can't keep up with him", speeds.

There are sooo many POS blocks coming in now... that they are killing about 50% of the attempted postings of POW blocks...

Enough that the diff is about half what it should be.

Please set the diff high for POS, until this is fixed, to stop those still mining POS from completely degrading normal mining. A few are still mining POS and getting tons of them faster now. Only having one with POS mining simply makes that one get all of them. It does not slow them down at the easy diff that it is set to.

I assume they are not being booted, because apparently there is only a few, or only one... as opposed to the massive collisions when most of us were still mining them.

I think we can all agree to that. Then we can all start mining POS again, and you can manually lower diff with each version. (That should give you an idea what to set for a diff adjustment for POS work. Even if it is something simple like 5% of POW diff + 0.02.)

Normally at 100K (1.53 diff), I get about 2-4 blocks an hour... Normally about 10% rejects.

Now, at the 40K (0.61 diff), I am getting about 1 every two hours, due to collisions {rejections}. Now about 75% rejects.

And the diff keeps falling to attempt to compensate, creating more POW to throw and try to beat the POS being generated. (There is no balance)

Speaking of limits, you might want to put a lower-limit on POW... like 45K (0.68 diff), which seems to be a choking point for most "fast coins" if they go lower. More collisions/rejects.

It is suggested to... (If you know how.)
1: Ensure port forwarding is setup for your router/modem for port 7685 (TCP access)
2: Possibly ensure, if your firewall is on, that bottlecaps-qt is set to "allowed"
3: Add this line to your config file...
reservebalance=1000000
That will stop POS blocks from being created, for now. As long as you have under 1,000,000 coins
4: Follow the block-crawler height, which seems to be the "tallest chain" always.
http://bottlecaps.kicks-ass.net/block_crawler.php
5: Also, from that page, use the "Connections" link to get the latest peers to connect to, adding them to your config and deleting your "peers.dat" before you restart after being locked-out.
http://bottlecaps.kicks-ass.net/peers.php

If you get locked out...
1: Turn off your miners and shut-down the bottlecap client.
2: Delete the bad chain data.
FILES: ".lock", "blk0001.dat", "blkindex.dat", "db.log", "debug.log", "peers.dat"
and the FOLDER: "database"
3: Restore those files from your backup, if you have one. (If not, just start your wallet and rebuild it from scratch.)
4: Wait for a moment while the wallet pokes through the rooms, and connections, looking for the tallest chain, while it builds your database to the best height. If it does not match the block-crawler, it might be the wrong one, unless you found one that is taller than block-crawlers height.

To backup...
1: Wait until you are connected to the tallest chain.
2: Create a folder called "backup"...
3: Copy and Paste these items into that folder...
FILES: ".lock", "blk0001.dat", "blkindex.dat", "db.log", "debug.log", "peers.dat" (Not this file)
and the FOLDER: "database"
4: Use those backed-up files to replace a screwed-up database that is stuck on the wrong fork, at the wrong height. Follow the instructions above for "If you get locked-out..."

Optionally: Backup your wallet every day... Just in-case of complete network nuclear melt-down!

Other things to think of using...
From the debug window console. Found in the menu-item "Help".
- Type "repairwallet", if you just left a bad chain. That should restore some lost coins that may seem "missing in action", due to POS invalid blocks, or POW missing blocks.
- Type "getpeerinfo", to see a list of connected peers and information about connections.

If you have time to babysit the machines... I say keep solo-mining (If you have that hash-power.)

If not... stay on the largest pool that "is the big chain", and will always remain that way, until resolution comes... Which I believe is verns pool. (But don't quote me on that!)

It won't be turned-on in cryptsy until it has been resolved, and without issue for more than 24 hours... That is what I was told last night. It has not been 24 hours without issue yet. (Only for those who have stopped POS mining and had a direct connection to that tallest chain, without having to depend on another connection to act as a relay to that taller chain.)


That was one of the times I was severed also... then again at about 3:45...

Thanks for the compliment...

Still a few people out there aiding to the problem, instead of assisting the resolution. Make an hourly check to ensure you are still mining on a live chain, and eventually, if POS is disabled, you should get a direct connection to the tall chain that does not seem to be mining POS, and continues to thrive.

Plus, those POS blocks seem to be causing a lot of rejects, since they are found almost instantly... invalidating legitimate POW blocks that just can't be sent fast enough to be accepted over the faster POS, causing us to start all over again, looking for a new hash-solution to the new block-height and new POS hash.

Might want to consider making a version 1.4.3 that does not have the ability to mine POS and rejects any POS submissions, pending the release of 1.4.4 with the "fixed POS" issue. For those who stubbornly are keeping POS alive, for those ten pennies a day.

If you are pool mining, then please turn off the wallet so it does not POS mine... You will eventually end-up filling your wallet with hundreds of invalid POS blocks that will eat-up your coins, until they get returned from the wallet, when they expire.

If you are mining POS and also attempting to be a node... you are getting anyone connected to you banned and making them isolated, as they mine off your invalid POS blocks, when you get severed from the chain. (If you get severed from the chain.)

At the moment, there is no reason to mine POS, as I do not believe it is rewarding or functioning properly anyways, from what I saw when I was mining POS.

If your wallet was banned, it would reconnect again, if a new miner connected, and acted as a relay/bridge from you to the taller chain, but only if you delete the old chain and restart and find that miner... Otherwise, you will just keep connecting to the wrong chain, continuing where you left-off, on the wrong forked block-height.

If you want to help the network by keeping your wallet open as a node, then please don't unlock it, or set the reserve high enough so it does not attempt to POS mine.

Until all of this is resolved. (Will not be difficult to tell when it is resolved, some people are persistently ignoring the request and mining POS, and ending-up alone and taking others down with them. I am sure we will hear when they have fixed the issue, because they will stop posting here about not having connections or getting booted.)

Add this line to your config...

reservebalance=1000000

That should stop POS creation for you, making it possible that you MAY live longer... The POS 0.1 coin gain at this point is not worth the 100 coins you would have had, if you were to stay connected longer, mining only POW blocks.

Just hope that the person who is connecting to you, is not generating POS that will get you banned in the process for relaying that POS for them, to the chain. Or that you are not connected to someone who is generating POS blocks, and will be banned from the larger chain. Which will leave you blindly mining on the chain you believe is valid, because you all don't know the other chain still exists. Because the other chain is not accepting your connections anymore. Apparently, that is what is killing me. Guilt by association.

I took the time to reconnect again, and will just babysit this bad-boy... Obviously "Vault 108" is still active... Google it. lol. "Damn you Gary 54!"

It is not being attacked, that I know of... Seems to be self-destructive, due to an issue with POS.

Something about submitting undesired POS blocks... apparently multiple times, which is causing the network to kill connections. But instead of killing only one connection, everyone attached to that connection also seems to be getting banned for "misbehaving". Leaving us talking only to one another, and building our own "valid" chains that become a fork. With no way to reconnect to the other fork, because they continue to ignore us.

Eg, technical difficulties...

I am sure other POS coins had similar issues... Might be worth checking-out how they handled it, instead of just comparing this coin to one other coin. But I suspect it is one of the values related to difficulty and distribution and allowance of resending, being the major issue.

EG... You submit a POS or POW block... At that time, the network you are on already has +2 over the height where you submitted your block, but they just didn't relay that info to you yet... Now it looks like you are misbehaving, sending a POS when all POS for this time have been used-up, and you resubmit it six times, as you are not getting feedback saying "No, stop sending, there is a new height"... Thinking they just didn't hear you, or it got lost in the send... POOF, instant ban... Now you and the three people who did hear you, and also didn't get the new height yet, simply accept yours as valid, and they stack three more POW blocks or POS blocks, and you happily mine away on your own private chain/fork... Since the other taller chain which is now 7 blocks over you, isn't talking to you, and letting you know they are taller.

Eventually you end-up with one fork at a height that is less than the one with the most hashing power. Which always becomes the tallest chain, after it has stripped itself of all but the three required minimum connections. (Eg, 2 more newbs that have not yet been banned join, keeping that tall chain alive, until it allows us to reconnect again, an hour later.)

This is where having a centralized component would aid in the issue. If anything, it could simply let everyone on the shorter chains know NOT TO MINE, and force them to attempt to find the larger chain, or even relay what the nodes are refusing to relay... the blocks... Thus, stopping them from dominating simply because they are the largest single entity. (Because if another smaller miner submits a block height larger than the one they had, they would have to wait to get it. Pulling it from the centralized node that is just a node, not a miner.)

Happened again at 5AM... I surrender...

Waiting for version 1.4.3...

Sitting here "not mining" on a "locked wallet", I remain stuck at block 80811 with two connections, and all other connections rejecting my connection.

It can't move to the higher chain, because they are not even telling me there is a higher chain, before they disconnect. That also needs to change, even if you are "banned" or whatever... The least they could do is tell you there is another height, so you stop building one alone. So these morons that are the only ones I can connect to, who are obviously part of the issue... don't let me connect, because they would know that they are not on the right chain also, and stop adding to the problem.

Signed, Really frustrated now.

lol.

Port is now forwarded... but nothing has changed, except I still have 2 at a time and 2 trying to connect, then they disconnect from me.

Still only have 2 connections that stay... Had 4 for a moment, but lost them with port forwarding or because they got bored with my posts and left for psychological reasons of keeping sanity. lol. Bounces from 2-4... but only 2 ever stay. The same two stay...

Again, just to remind you... Just before 1AM, I instantly had 8 solid connections that stayed. It wasn't until the fork that died, which left me mining with 2 connections, then 1, then none... until 3:30AM, when this stuff started. So I doubt it is "me" or "my settings", as nothing has changed.

This is what always happens... it starts out perfect, then once this happens, it essentially kills my ability to mine. It took a lot longer this time, but it still happened. It will eventually kill me again, I am sure, just after I lay down for sleep again.

Also, if it matters, I am in the USA, Florida... on high-speed internet cable-modem att-uverse... and now, directly port-forwarded for this computer.