Quote from: phrozenspite on March 08, 2014, 11:04:07 AM

Quote from: poolwaffle on March 08, 2014, 10:23:37 AM

Quote from: xia0mingx on March 08, 2014, 04:26:31 AM

Hi,

When everything stablise, will there be an Asia server for WafflePool?

Thanks

Yep, its actually up and running, just not in DNS yet. Moved USE and EU to new servers late last night, and then my internet went down (working off a tethered 3g connection now, hooray). Going to move USW over to the new servers, and then make sure SEA is up and running. I'll post back (and put it on the site).

Quote from: richmke on March 08, 2014, 04:35:59 AM

The next time you are doing something that will take the servers off-line, can you please post a message on the Stats page? Advanced warning would be nice so I can point my miners somewhere else. Lost 8 hours of mining because I could not reset my miners when I am away from home.

I mentioned it multiple times in the thread, as well as on the home page for a few days (that there might be a tiny bit of downtime and to setup a failover pool). When it was actually changed over, there wasn't more than a couple hours of up/down time (where servers were constantly switching), and during that time it was more a matter of just disconnect/reconnects as connections were switched to the new servers and back to the old, definitely shouldn't be 8 hours of mining time (from start to finish of the full switchover was about 4).

That said, a backup pool is _always_ a good idea. In case we get DDOS'd, or we have emergency upgrades (or planned upgrades), its good to have your miners automatically jump to a working set of servers (either another cluster of ours, or somewhere else).

How are the new servers looking so far?

Definitely better. That said, not as good as I was hoping

Adding SEA might help (some users will move from eu/usw to sea). I was hoping the new servers would give us room to potentially double capacity, but it doesn't look like we got that. First estimates are maybe +50%, instead of +100%, meaning theres some limits we're running into that are a bit tougher to track down (potentially network bandwidth, latency, buffer sizes, kernel parameters, etc).

Once these are up and good, I'm going to start working on optimizing the actual stratum servers a bit further (read: rewriting pieces). We shouldn't be capping out this large of boxes at the current loads, which normally means poorly written code somewhere