downtime noticed
 

The Cellar was offline from apx 3:30am to apx 10:30am EST due to a system crash. Root cause not identified, but it happened at the moment when system logs are rotated and processes restarted, and there happened to be an unusually large log file in the mix (no malice involved, just a log I hadn't rotated before).

I've taken this opportunity to upgrade the system as our host began offering double your system memory for free. (This actually may eliminate the source of the problem.)

This means Cellar does not qualify this year for the "five nines" category of "highly available", and only qualifies as "three nines". (It's only available 99.9% of the time, not 99.999% of the time.) In recent years, the regular failure of platforms like Amazon AWS has meant that even huge sites have had a hard time managing to reach that goal. So we're doing OK.

http://cellar.org/2017/99999.jpg

No single-point-of-failure solution is ever really highly available. To qualify we would need failover with at least one other server in a load-balancing platform.

We don't need the cover story, we're just glad you changed your mind.

I blame xoB -when it can back up the list of new posts was all him, at that time.....

I'll take the blame, I have a broad's shoulders.
I'll admit I was late with my shtick yesterday.

Looks like I got here just in time. Late.:D