Jump to content

Obsolete:Labs Eqiad Migration/2014-02-26

From Wikitech

Present: Mark, Andrew, Diederik, Ryan, Coren

Feb 27


Andrew wonders about:

Shared storage migration: 

   - root between instances?
   - pre-emptive root migration?
   - use sparse rsync
   - total usage shared space 7.5 Tb
   - 12Tb for /data/projects
   - we expect to copy at least a couple of  100Mb/s
   - ACTION ITEM: verify speed copy pmtpa <-> eqiad
   -

Virt1000 web interface issues: 

   - integrate DCs on wikitech?
   - Where to update instance info?
   - Keystone is ready to be deployed -- need some more testing of eqiad instance creation
   - ACTION ITEM: Ryan will add to Keystone group specific rights to instance creation

/var/scratch failures on migrated instances

Need to write migration guides for volunteers

  • Andrew will give this to a shot in the coming days

Create bugzilla queue for migration requests

  • Separate component? Diederik can talk with Andre Klapper (ACTION ITEM)


Status action items: 

   * Figure out how not to NAT to other labs cloud (Mark) (DONE)
   * Run salt to detect self-puppetmaster merge conflicts. (Coren)
   * Estimate time it takes to migrate instance (Andrew) -- tomorrow some first numbers don't forget to add sparse file option
   * Work on keystone to share keys between tampa and eqiad. (Ryan) DONE
Retrieved from "https://wikitech.wikimedia.org/w/index.php?title=Obsolete:Labs_Eqiad_Migration/2014-02-26&oldid=618264"