Microsoft brings Azure back online
Red-faced Redmond mends worldwide SSL certificate cockup
Ensure Ease of Recovery with Asigra’s Agentless Software
Microsoft has managed to repair its Windows Azure cloud, after an expired SSL certificate downed storage and other services for people across the world.
Ninety-nine percent of the affected services have been brought back online, Redmond said early in the hours of Saturday morning, Pacific Time.
"We will continue monitoring the health of the Storage service and SSL traffic for the next 24 hours," the company wrote in a post to the Windows Azure Services Dashboard. "Customers may experience intermittent failures during this period."
The global outage lasted for around 12 hours and occurred because Microsoft failed to renew a security certificate.
The storage knockout created a cascading series of failures in other crucial parts of Azure, eventually bringing down Xbox Live components as well.
In the coming days, Microsoft will have to explain to its customers how a crucial security certificate was able to expire and bring down all secure storage services around the world simultaneously.
Some customers may be hard to mollify: a certificate problem was also to blame for a widespread eight-hour Azure outage in February of last year.
Technology industry insiders expressed a mix of disbelief and sympathy for Microsoft, with many asking how it was possible for Redmond to fall foul of such a basic error.
Netflix, a major customer of Azure's cloud rival Amazon Web Services, developed a software package around two years ago called the "Netflix Security Monkey", which automatically checked certificates on the company's cloud to avoid failures like this, its chief technology officer Adrian Cockcroft told The Register via Twitter.
The Reg's cloud desk will follow the aftermath of this story with interest, and is keen to hear Microsoft's explanation for the Blue Sky of Death. ®
COMMENTS
Cron job needed and, er, why didn't they renew the cert for longer?
OK, so this happened last year, presumably on the annual renewal date. It begs how incompetent Microsoft is:
1. Most secure cert registrars send out e-mail reminders (mine does with 90, 30 and 7 days to go) - did whoever they registered with not send such e-mails or did Microsoft just ignore them?
2. A simple cron job to check the cert and e-mail (to more than one person!) every day at least 7 days before expiry would have saved their bacon.
3. When they messed up last year, why didn't they renew the cert for more than one year? Surely Microsoft can afford a multi-year cert?!
Multiple levels of incompetence there - that's Microsoft for you.
Microsofts "customers" get Ballmered again
Having stuffed them on phones, tablets and their comedy new desktop OS, destroying their faith in their cloud offering is an essential step in the slow car crash suicide that is 2013: Microsofts year of hell.
@Eaton -> Just stop it.
I am dealing right now with a domain that has been hacked three times in the space of a week, on a Linux server running Apache. I have to recreate the site because the fine technical folks at this Linux-based host overwrote the backups with the hacker's new site design. Yeah.
I'm currently going through WayBack to get an idea on the site layout (designer lost original files and the webguy who put it together literally went crazy) and found the vulnerability, a known issue with Joomla 1.5 that the host never updated.
Now, do you really want to discuss this? Damn fools and idiots are everywhere and the use all kinds of software -> the discontinuity lies between keyboard and chair, buddy.
IT infrastructure monitoring strategies
Agentless Backup is Not a Myth
Top 10 SIEM implementer’s checklist
Steps to Take Before Choosing a Business Continuity Partner
Enabling efficient data center monitoring
