Power outage in Linode Fremont Datacenter | Hacker News Reader

Power outage in Linode Fremont Datacenter

status.linode.com

39 pointsdrm23763 comments

My linode has been down for over an hour and still counting. They just went under three nines in one day...

I noticed this an hour later because my monitor script was set up in the same datacenter. Lesson learned.

I have zero affiliation with the group, apart from being a casual user but http://wasitup.com/ has come in handy for me quite often. Even for free, it'll do a check 20 times an hour.

Pingdom also has a very good free plan. It's limited to one url, but for that URL, it's just as good as their paid plans. So you get checks from a ton of DC's, and it tracks response time as well as uptime.

It hasn't been well publicized after the original announcement (and isn't prominently placed on their plans page, either) but I've had great success with Cloudkick's free developer plan on my personal virt (which is still down..):

https://www.cloudkick.com/accounts/signup/developer/

CK sent me a text message as soon as my host went down earlier, which gave me time to file the ticket against Linode.

Only caveat - they don't provide a signed RPM, which can cause some administrative hardship for linux users using rpm-based distributions (most of the yum operations require you to add an extra cli option to allow working with unsigned rpm's), but it's a minor complaint given that their agent isn't mandatory and doesn't need a lot of updates.

Another vote for wasitup.com. I use wasitup.com, host-tracker.com, and basicstate.com, and wasitup.com is consistently the first one to report an outage, followed by basicstate, which has a surprisingly gloomy and odd interface, yet good service, and then followed by host-tracker.com, which sucks pretty much on all levels. All are free btw.

Just keep in mind that I host http://wasitup.com in Linode's London facilities. Data center redundancy has been on my TODO list for some time now.

So... they ever heard of redundant power generators? Does anyone know what happened exactly?

I've had power transfer switches fail more than once. Just because you have a redundancy available doesn't mean you are reducing failure points in the system.

I live right down the street from the datacenter (~ 0.5 miles). The whole area's power was out for a bit, but know more than an hour. My site is having errors.

I had just brought up a new instance. Why not choose Fremont? It's right down the .... oh dammit.

That's a good case in point. In general you're better off choosing locations further away from where you live/work to prevent issues like this where your office/home and datacenter could be affected at the same time.

It was especially dumb of me because my other instance was ... in Fremont.

We had just launched a test server about 40 minutes ago and had been waiting for it to boot for ages when I finally thought to check the status page. Talk about bad timing.

Still wait for the reboot jobs. maybe I should move to EC2.

Yeah .... our iPhone app downloads content from our server - so far we've got 3 1-star reviews saying 'Your app sucks - it can't download any content'

Damn .........

my server is back.

http://aws.amazon.com/ec2-sla/

Their SLA says 99.95% uptime. What has your Linode uptime been?

The Fremont datacenter, if I remember correctly, is at Hurricane Electric. They had power issues in 2009 as well.

http://blog.feedly.com/2009/11/03/24-minutes-of-unscheduled-...