Past Incidents

There was an issue on Graham, which is still being investigated, between 8:09 to 8:19 that caused job failures. User's would have received a message with an error stating "NODE_FAIL". If your job was affected by this failure, please re-submit it. Sorry for any inconveniences this has caused.

Cedar Outage Oct. 15-19 The Cedar system will be unavailable during the week of Oct. 15-19 for operating system updates, file system upgrade and modifications. Jobs that are still running on Oct. 15, 8:00 (Pacific) will need to be terminated.

FIXED: the problem was only of connectivity, and running VMs were unaffected.

The planned power outage at the UW datacenter has unexpectedly affected Graham Cloud services. This is affecting network connectivity to VMs, and may affect them more seriously (such as storage connectivity).