Simon Willison’s Weblog

App Engine outage postmortem. Interesting peek behind the scenes. The primary cause of the error was a bug in a GFS (Google File System) Master server caused by a MapReduce process sending a malformed filehandle, reminiscent of the error which took down S3 last year.

# 9th July 2009, 12:49 pm / appengine, downtime, gfs, google, s3

Kosmos Distributed File System (via) New open source distributed filesystem similar to Google’s GFS.

# 28th September 2007, 9:12 am / gfs, goggle, kfs, open-source, richskrenta