Hacker News Reader

Comment by _hnwo | Hacker News Reader

Now this is how you do outage responses & notifications.

I'm not even a Linode customer but this kind of clarity and response is exactly what I want to see from my actual service providers.

View on HN

No, it's not. A fair bit about this sucked. I found out my linode was down from Cloudkick... tried to ssh, no go. Logged into Linode.. My dashboard says that everything's peachy and my Linode is running. I start freaking out thinking maybe I ran it out of disk space and it's thrashing so I start issuing reboot requests.. safe-mode-reboot requests, etc.. They stack up. Nothing's being processed. Can't console-in either.

Finally I contact support. They were VERY fast in answering, and when they did they pointed me at status post that said Freemont is kinda out...

WELL THEN WHY THE FUCK DOES MY DASHBOARD SAY MY INSTANCE IS RUNNING?

Lame. Oh well. The 45 minutes or whatever they were offline had another 10 minutes of strapped-CPU reboot thrashing to get through before it was back up.

Oh, and this is the third time this year at least Freemont's been down.

So yeah.. cheers for being nice, and whatever.. but I'm not suuper enamored with how much love HN gives folks for ridiculous fuck-ups followed by a "Hey guys, I'm really sorry but here's a long blog post about how our flux capacitor won't ever fail us again (until next month)" blog post.

[Edit: Yeah.. I know. Downvotes were expected. Sorry for the rant, but this little outage of theirs caused a pretty embarrassing situation with an important client. So I'm happy to trade some karma for the opportunity to say 'fuck' in all caps in a public place right now. Cheers.]

[Edit: Oh, the irony! I post a totally obnoxious comment who's valid-points-to-tantrum ratio is way off the tantrum scale... Downvotes immediately. Then I edit a little (heartfelt) apology into it, and upvotes away. This is exactly the sort of psychology that's got me at least a little bit confused and intrigued about HN, and that I was tantruming on about in the original comment.]

You should have just jumped on IRC prior to panicking--you would have gotten an update instantaneously. Or better yet, you could idle a detached screen in there. When I had a Linode a few months back, there were constantly 500 rather helpful people in their channel.

Dude, I should have done LOTS of things (including not posting the above comment).. But the fact that their own site was saying "Everything's Clever!" really led to believe that I was the one that effed up... and I wasn't thinking clearly. My wife had literally delivered a big job to a client moments before so I was just in a hot panic to get it up.

Really, if there's a problem in Freemont.. and I'm on my dashboard for my Freemont Linode.. there ought to be a banner or something saying "Hey buddy, it's probly us this time."

Your absolutely right about the dashboard, but does it give a notice if the whole data center is having problems instead of parts of it?

Yeeah, that bit about where I said I wasn't a customer of them.

So, I'm mostly just talking about the responses I see on the status page and their follow up.

As a systems dude I prefer seeing a detailed response after the incident rather than excuses.

Reply on news.ycombinator.com