Failure is always an option

  • 12 March 2011

Blogging has been a bit neglected lately. This is due to the influence of crappy software and non-existent designs of Things™. Let’s get started with the Cloud. Once upon a day a server disappeared. Investigation yielded: The ISP has a problem and will fix it ASAP. The monitoring system kept sending alerts for two days. Further investigation yielded:

  • The ISP had fixed the problem. It involved a RAID controller and a filesystem, both probably corrupt.
  • Some (virtual) servers died in the process.
  • The ISP advised to recreated the deceased system with a fresh installation and backups.
  • We had backups. Most other Cloud Believers have not. Tough luck.

Then there’s IPv6. It runs smoothly in two local networks (these networks being not local anymore, NAT has to die). Apart from some applications which don’t get it everything’s ok with IPv6. Suddenly my ISP fiddles with the modem’s firmware, IPv6 at the other end of my last mile appear briefly, then they are gone again. This incident has left the MTU between my VPN server and me slightly changed, so that IPv6 tunnelling stalls. Reset the MTU to 1280, all works again.

Then there’s…ah, crap. I don’t care. I’m tired. Good Night.

Sorry, the comment form is now closed.