This should be fixed now.  I have quite a few security measures in place to make sure BTC isn't paid out multiple times.  When an attack like today's happens, sometimes those measures don't get properly reset on accounts that were active during the attack.


For Stratum users, if you're not seeing shares show up on your workers, try restarting the miner (or stratum proxy).  All servers should be working now, though the getwork servers may have some occasional delayed work due to one server still being offline.

I'm putting up some temporary servers to pickup the DNS entries that were nullrouted.  Most DNS entries still work, except the 'default' most people use, regular btcguild.com .  Using any of the other DNS entries should work normally (mine1, mine2, de, mergedmining, etc.).  Regular btcguild.com may be somewhat laggy for getwork based users.  Stratum users shouldn't worry since the getwork server will send you an IP to connect to, which re-establishes the connection rather than being proxied.

Another IP has been nulled, the website is back up and responding.  I'd recommend using the de or mergedmining servers at this time, since the attack is ongoing.  I will not be moving DNS again until I believe it is safe.

Contacting the DC again to null another IP.  Sorry again, but there's not much you can do when your server is being flooded for 6-10 (best range the DC has given me so far) gbps on a 2 gbps uplink.  The switch has no option but to drop 66%+ of the packets.

Fixing DNS for those that can't reconnect.  Hopefully your DNS isn't cached too long.

Everything is back up and running smoothly.  I will be going through and removing payouts that were recorded but not properly made (due to bitcoind timeouts).

8:14 AM (PST): Just got off the phone with support at the datacenter, they're going to take a look at the switch and see if they can diagnose the problem.  If not they will be performing a factory reset to see if that solves it.
8:55 AM (PST): Support responded.  This was a DDoS attack, not a switch failure.  The attack has been filtered upstream and service should be returning shortly.
9:05 AM (PST): Servers are coming back online.
9:15 AM (PST): Pool servers are back online.  Website coming up shortly.

[Website Status]:  Online
[Database Status]:  Online
[US-Getwork Status]:  2/3 Online as normal, other server DNS has changed
[US-Stratum Status]:  Online
[MergedMining Status]:  Online
[Germany-Getwork Status]:  Online
[Germany-Stratum Status]:  Online

Doing a complete power cycle on the entire rack in Chicago.  Hopefully things will be back online shortly.

I am -very- sorry I did not get on this faster.  Whatever is happening to the servers is not triggering my alarms that should have caused my phone to wake me up.  Working on fixing this ASAP.

Merged mining server is being restarted to fix a small problem on the server.  Since the mergedmining.btcguild.com server was initially a testbed for Stratum, it was not configured to allow large numbers of connections.  It recently hit the file descriptor limit.  Should be back up in a few minutes.


UPDATE:  The server is back up and running.  Sorry it took longer than expected.

An automatic payout that ran yesterday did not properly complete, resulting in a few payouts never hitting their destination.  I've removed this broken payout from the history and it will be reprocessed shortly.


The December 19th update is going to end up pushed back to December 22-23.  Christmas always seems to cause delays in my plans .

Everything is back online.  Appears it was some internal routing issue with the datacenter.  Total downtime was ~10 minutes.

Yes, there's heavy packet loss to the server that started up about 8 minutes ago.  I'm not seeing any signs that this is an attack, but it's very unusual.  The longest outage I've ever had at the datacenter is ~5 minutes in the last year, everything else was either a DDoS or my servers acting up.

Dear god, I can't imagine how much whining I'd hear if I tried to get people to move to Stratum like that... I should try it the next time I'm feeling sadistic.

I have a nice annoying message to anybody that hasn't used Stratum in the last 2 weeks.  They just ignore it.  I guess I can make the font size increase by 1% per day ;p.

77%?!  And here I am getting happy when users on Stratum have speed exceeding 50% (which only happens once in a blue moon).

As for running a dozen 60 GH/s miners.  Assuming we ever get to that point where a single device is registered as 60 GH/s [something tells me those ASICs will be recognized as multiple smaller devices], and somebody has a farm like that with a single controller PC, the miner could use ntime to reduce overhead if needed.

Assign each device ntime+x, and send them the same work with a unique ntime [this is assuming the devices do not support internal ntime adjustment].

I will be pushing an update to the Stratum server on December 19th.  This change will allow you to set a minimum difficulty on any worker.  Here is how it will work:

1) Difficulty will be adjustable under a new settings page (may involve some changes to Manage Workers).
2) When authorizing a worker with the Stratum server, it will load the difficulty settings you've set.
3) If the newly authorized worker is set to a higher difficulty than your connection is currently set to, the difficulty will be increased immediately.
4) Your difficulty on that connection will never drop below the minimum difficulty of that worker.


For most configurations (native stratum miners), this will be straightforward.  This can get confusing for a Stratum Proxy, since difficulty is set per connection, not per worker.  If you authorize many workers through one Stratum proxy, your difficulty will be set to the highest authorized worker's value.


Var-diff will still adjust difficulty higher if you are submitting shares too frequently.  Recently Stratum pools were updated to a much higher Shares-per-Minute setting, meaning it doesn't adjust a worker unless they submit more than 40 shares per minute.  This reduces overall variance, at the cost of slightly higher bandwidth.  This is still drastically lower than anything you would experience using getwork, and still ends up being only ~0.1 KB/sec data transfer.



This notice is going up almost a week in advance to remind miners that they should set backup pools.  You can set a different BTC Guild server as your backup, which will keep you hashing with BTC Guild during the downtime.  Ideally, your primary would be set to either btcguild.com or de.btcguild.com, and your 1st backup would be set to the other server.

I've always recommended automatic payouts be set to whichever of the following is smaller:

1) 1.0 BTC
2) 1 week's worth of mining proceeds


This will help keep your wallet "clean" of small transactions, which make it easier to spend your coins without having to pay any fees.  This also means you're not trusting the pool operator as a bank, and at most they will only be holding ~1 BTC of your money at any given time.

Most automatic payout pools process multiple times per day, so these settings work for the majority of miners.  However, if you're producing 24+ BTC/day, you might want to set it slightly higher, since I think the majority of pools only run automatic payouts hourly (or less frequently in some cases).