Craig Ulmer


2014-02-06 clusters

Whoops. A couple days ago someone pointed out that our cluster's login node was down. Earlier in the week we'd asked the data center guys to upgrade the hard drives in an unused section of the system, so I figured someone must have powered something off and forgotten to turn it back on. On inspection today, we discovered that instead of upgrading 8 computers, the workers had managed to get 9 done. Unfortunately, that 9th one happened to be our login node (the only thing that was supposed to stay live during all the upgrades). Double unfortunately, our data center's policy is to physically obliterate used hard drives once they're pulled, so there's no going back. On the bright side, I guess the login node is going to have a lot more space once I get it rebuilt (a shame this one wasn't a oneSIS).