Good morning fellow data hoarders. Yes I admit it I am one as well!
We’re still seeing lots of very heavy IO on the Archive side, we honestly weren’t ready for the inrush of customers at the very last minute. Things are easing off however and nothing (neither bandwidth nor disk IO) is at the moment a limiting factor. We are using spare IO bandwidth on our back end now to re balance clusters.
We’re making plans to improve things before we offer the sale again. Part of these plans includes a 10x increase in Archive’s bandwidth (We knew we would need to do this to offer what is coming next!). Starting on June 1st Archive will increase per-customer upload speeds from 300Mbit to 1Gbit PER CUSTOMER :-).
Lastly we are doing some back end work on 3/13. Three of our storage network’s Ethernet links have been persistently dropping out. So much so that we’ve the removed the entire port channel from use. This takes some 30Gbit of capacity off the table between edge and our data stores. We want this back! Our working theory is bad cabling and we’ll have new cables ran via a different and shorter path this weekend with the goal of turning them up on 3/13. This SHOULD be seamless but the key word is should. So heads up there MAY be a brief service interruption early in the morning on 3/13.