Website got a quick burst.  The current website reverse proxy is a bit more vulnerable to attacks now that we're not on Cloudflare.  Should be working again (wasn't down very long).  Backends are still mining just fine.

Should be fixed now.  There was a random issue with certain JS failing to load due to some optimizations I added to reduce overall page load time.  Recently added a more complete fix to the issue, so on a refresh you should have all the charts working again.

Just FYI:  There is no "static diff" on BTC Guild.  There is a *minimum difficulty* setting, which starts your miner off at your minimum and won't adjust it below that.  Vardiff will always adjust you *higher* if your hardware is pushing out too many shares.

I guess that is another factor.  If the units internal reporting and BTC Gulid's dashboard are fairly close after an hour, it may be that these rejects aren't even "wasted" work.  I know my old ztex FPGAs used to have a problem with duplicate shares, anywhere from 8 to 10%.  However, they weren't doing duplicate work, it was just a bug in the firmware not clearing out the "winning nonce" part of memory when it submitted the share, so it would resubmit it each time it finished a full nonce range without a new result [pretty sure that's what it was at least].  The dashboard/payouts were matching the speed reported by the mining software.

Strange, considering I know quite a few KNC units that mine on BTC Guild since day 1, and have only just heard this now.  Are you two running the latest firmware?  Was it doing this *before* the latest firmware?  Multiple pools configured?

I didn't "have it" unless you're talking about 6 months ago.  BTC Guild hasn't been above 32% of network hash rate since ASICMINER left. Hell, at this point referring to any public pool as a threat is losing sight of the larger picture.  GHash.IO will pretty soon be the largest amount of network hash rate (now at 440TH/s, a massive jump in almost no time at all).  ASICMINER is supposed to be putting up 500TH/s in the next 2 months as well.  KnC is starting their own private pool as well.  In about 3-6 months, you're probably looking at all the public pools *combined* not even representing 40% of the network.

What miner are you using, it's SEVERELY broken.  You're submitting 500k+ shares that didn't come from BTC Guild.  Your 500k+ "Other" shares are Unknown Work, meaning you're returning results to work that was never given to you, either because your miner is spitting out corrupted data/unpadded results, or submitting work from a different pool.

Except BTC Guild hasn't been 40% for roughly 6 months.  Maybe you should learn more about the actual network hashrate distribution instead of quoting shit from reddit/forums from blockchain.info's 24-hour charts.  BTC Guild was only ONCE above 40%, and that was entirely due to ASICMINER not being setup for solo mining when they first started up.

There were some reports from users with bfgminer using startup scripts that immediately connect to multiple pools NOT using failover-only getting messages related to the pool giving them old/stale blocks on stratum.btcguild.com .  Last night I pushed out an update to the code on that server which hopefully fixes the problem.  This error was related to how the server was changed to filter/validate miners to separate botnets/DDoS traffic from legitimate users.

If you were having this problem previously, please let me know if it is still happening.  I have not patched the eu-stratum.btcguild.com server with this update since I want to make sure it works first, but I have so far been unable to duplicate the issue myself.

Glad to hear it .  The reverse proxy was setup to cache all the static files locally, but it wasn't redirecting the CSS imports properly so outside of the main CSS file nothing was styled.  I changed the links slightly so your browser *should* automatically get the corrected version now even if you had the bad version cached.

Should be fixed now.  If the page has virtually no styling to it, clear your cache (or hold CTRL and hit F5 to do a force reload).

The website speed should see a decent speed boost once your DNS updates.  Last night at the start of the server move, an Amazon EC2 small instance was being used with iptables to simply proxy requests to the real server (keep it protected from attack) to get off of Cloudflare ASAP.  Now migrating DNS to a dedicated box using a more sophisticated reverse proxy setup to speed up the general navigation.  It was never obscenely slow, but there was definitely a minimum ~150ms delay added to each page load due to the EC2 proxy.  The new server should reduce that delay by a reasonable margin.

This should not trigger SSL complaints like the previous switch.  The new reverse proxy responds with the same hostname and SSL certificate as before.  Hopefully by now most/all users are no longer seeing a Cloudflare issued certificate due to the slowness of nameserver updating.

Hah.  Luckily to date no security fix has been vital, it's mostly about limiting the effectiveness of brute forcing by either eliminating methods to reduce entropy (finding usernames via password reset attempts) or eliminating the ability to use a single machine for more than a dozen or so attempts.  Same thing with the recent changes to the Stratum frontend where it's validating legitimate miners over botnets/attackers.  It wasn't vital, but it severely impacted the efficiency of the future attacks as well as minimized the actual damage they did.

Well, it was bound to happen.  After months of an attacker proxying through MILLIONS of bots to get around IP bans, trying to find a collission of BTC Guild usernames/passwords from who knows how many database leaks from other websites, the attacker recently started trying a new vector.  In the last 12 hours, a significant number of password reset attempts were made, because the attacker is now trying to use the password reset function as a way to narrow their brute force attack down to only valid usernames.

New attack detection systems have been added now which have stopped the attacker from successfully determining if their attempts are valid usernames or not.  If you received an email about a password reset, this was likely because you either use a very common username, your username is publicly known (forum username for example), or your username is part of a database leak from another website.  However, knowing the username is valid does not mean they know your email, or your password.  Without being able to access your email, the password reset cannot be completed.

If you DID receive a password reset email, make sure your password is not reused from another website which may have had a compromised password list due to plaintext or poor salting.  If you received a reset email, your username was in a database that somebody has compiled which has passwords that your username has used on other websites in the past.

The problem is a significant number of those settings are completely ignored by ISP level DNS caching/resolvers.

Invalid certificate errors are expected for people who use DNS resolvers which are still using the cloudflare nameservers.  It's been about 18 hours since moving to a new nameserver, ideally you would be updated already.  It only took about 4 hours to update on my PC using Google's DNS servers.  There's absolutely nothing I can do on my end to fix the cloudflare certificates, it's just a waiting game for your DNS resolvers to use the proper nameservers.

DNS issues should be over now, I have moved the DNS off of Cloudflare due to their withholding the records for 'www.btcguild.com' from many DNS requests, effectively removing the site from the internet for random people using their ISP as the DNS resolver.  It may take a little longer for the nameserver change to propagate.

The site is moving a little bit slower as a result, mostly on the *first* page load, since you're now retrieving all the JS/CSS/images from the site instead of the cloudflare CDN.  Subsequent page loads aren't quite as quick as it was for me on Cloudflare, but are not too bad, at least coming from California (which is one of the slowest locations to access it from within the US).  I'll be working on tweaking things with the site this weekend to help speed it up.  Most of the slowdown is that I'm forced to proxy the site through another server in order to protect it from direct DDoS attacks.

Website isn't down, and it isn't under attack.  Cloudflare appears to be fucking up BIGTIME lately with ISP DNS providers.  People using Google DNS or OpenDNS for the DNS servers have no issues resolving the 'www.btcguild.com' subdomain, but a lot of ISPs are screwing it up.  I'm migrating off of Cloudflare tonight because they refuse to admit this is a problem on their end.  Going to miss the protection they offered the website, but their support is completely unacceptable when this issue is absolutely on their end given the number of users reporting it in the last few days.

0.00 Auto Payout = Disabled.  0.01 is the minimum payout in all forms on BTC Guild and that is not expected to change unless there is a *massive* increase in the value of BTC.  Manual payouts will include all the extra satoshis as well to allow a complete balance clear, as long as it's at least 0.01.

If BTC Guild merged mined those coins, it would be worth less than 0.02%.  Basically the only reason those are adding 0.2% is because you get *so* many of them per Bitcoin mined, due to no pool giving a damn about them.  The same thing happened with NMC.  It *used* to be worth a decent amount.  Hell, Namecoin basically died as a result of it and had to implement a hard fork (merged mining) in order to revive the currency to stop Bitcoin miners from hopping onto it when profitable and leaving after the diff change when it became massively unprofitable.

You'll also note that the extra % gained from NMC took a massive dive recently, because at it's peak it was ~5-6% (not counting the first month or two of merged mining).  Then BTC Guild added merged mining to Stratum.  Then ASICMINER did.  Then GHash.io did.  Now the NMC difficulty is > 50% of the BTC difficulty, but the price is still roughly the same.


EDIT:  Thank you however for reminding me that NMC is still 50 coins per block.  So my 0.6% figure was half of what the actual extra income is since I was calculating it without factoring in the 50/block reward.