In an interesting move to their new bigger better and more reliable data Los Angeles data center they managed to bring their site down as part of the cutover, synchronisation and testing. The official Linkedin blog advised that it should have been only offline for around 6 hours starting at 5pm on Sunday but at 8am Monday morning the site was still having some issues. It’s a bit of irony that the move of the production site to the more robust Linkedin data center lead to this downtime.
So what can be learned from this?
- If you have this planned please update your error message to reflect the likely issue is that the site is moving to a new server, don’t use generic messages and hide the story on your company blog.
- Linkedin was correct when they used a WordPress.com blog for their company news so any issues with their main site doesn’t prevent them getting their message across
- Linkedin did the right thing but it’s stream of updates via twitter but they could have also posted a quick note on their company blog
- Don’t self congratulate yourself before the move is complete, the staff jumped the gun early with this Twitter update
It’s great to see companies like Linkedin upgrading their infrastructure but others can always learn from their mistakes when it comes to your time to upgrade your servers and no Wikileaks is not responsible for this downtime.