wiki:DevOps-2014-03-04
Last modified 6 years ago Last modified on 03/04/14 10:49:03

DevOps Weekly Meeting | March 4, 2014

Time & Location: 9:30am-10:45am in LHL164

Attending

jpr, mhanby, tanthony

2014-03-04 Agenda

  • Login-01 & -02 functionality, network connectivity fabric
    • need vlan to login-02
  • CGHub tests
    • use nfs gateway for ceph container via login-02
    • need to run perfsonar toolkit and need to reconfig disk to support local storage
    • depends on vlan connectivity for login-02 to act as gateway
  • Detail RCS compute fabric "cheaha" and related services from BEC to 936 building

Summary

  • Will transfer zones from original environment to new locations.
    • Network gateway/fw (rcs-srv-01,-02) to rust
    • HPC admin (puppet, nagios, robinhood vm) to rust
    • Headnode to rust (on cloud-03)
    • Scratch (Lustre) - moved to 936 via front-end lustre nodes
      • About 65TB being used as USER_SCRATCH
      • About 20TB being used as SHARED_SCRATCH
      • About 50TB being used as lustre-recovered

  • HSV storage to RUST
    • Add

We need to notify the users that SCRATCH will go off line in May and that they need to move their data to their TB storage. LUSTRE will be reformated as part of the move so the data must be moved.

This depends on:

  • vlan sanity in superx
  • login-02 as stable network/file gateway (start calling it rcs-01)
  • ceph rbd volumes for $HOME
  • 720xd from hsv to rust and part of ceph
  • stage share_scratch on ceph?

More to come