Offsite Backups (old)

This page page may be outdated and could contain incorrect details. Please update it if you can.

Wikimedia is starting to use external partners to allow for better disaster recovery and quicker response time in case of a data center wide failure. The first large scale case use is being proxied thorough a company named eBart.

eBart

• Host: wm.arhiv.rs
• Port: 60222
• Alternate entry: wintermute.ebart.co.rs -> wm
• Primary Contact: Milos Rancic millosh@gmail.com
  • Phone: Buisness +38164822939 Personal +381656455674
• Secondary Contact: Branislav Jovanovic branej@gmail.com
  • Phone: +381604828520
• Tertiary Contact: Nikola Blagojevic nblagojevic@arhiv.rs
  • Phone: +381648220904
• User access: tomasz
• Type of access: ssh
• Space: 6.9T

The plan for this is to have a snapshot of Commons available offsite if anything should happen to our data center. The total allotted space should reach upwards of 10T's in the next months as more drives are added to the system. Ideally this off site will host both a snapshot of Commons and possibly an archive of the XML snapshots.

Notes on coverage:

• Between 10 and 17 CET/CEST (between 08-09 and 15-16 UTC) ability to reset the machine quickly.
• Between 07 and 01 CET/CEST (between 05-06 and 23-00 UTC) ability to contact other staff with a 20 minute window to reboot.

Status

All the 2008 archives are now moved over. --Tomasz Finc 18:10, 30 July 2009 (UTC)

Started to push [0-9]* & [a-z]* of Commons images --Tomasz Finc 20:35, 10 September 2009 (UTC)

The older archives from brion's drive (2002, some 2003, 2005 and 2006 dumps) are on the download server in a private directory and on a hard drive in a cabinet in Tampa. ArielGlenn 17:09, 20 January 2012 (UTC)

MIT

Samuel Klein is researching for dedicated space at the Harvard-MIT Data Center.

Stanford

Kul Wadha is exploring storage options at Stanford.

Amazon

We've pushed our initial data set into Amazon Public Data Sets and are awaiting legal and internal consistency sign offs.