[auton-users] AUTON Cluster Downtime Jan 4-8.

[Michael J. Baysek]

Hi Lab,

Auton / Neill Labs Central Computing System Downtime Announcement:

System will be down Jan 4th - 8th.

Detail:

The Auton Lab's central computing resources will be down for 4 full days
starting on Wednesday January 4th at 18:00 EST.  They will be back up on
Sunday January 8th.

The SCS Facilities folks are cutting power to the Wean Hall Machine Room
on Wednesday, January 4th at 18:00 EST.  This is the second of two planned
outages which will double the electrical capacity of the Wean Hall machine
room, and allow us to move some of our recently purchased equipment into
our server racks.

Power will be restored on Thursday January 5th, but I intend to keep the
system down until the Sunday the 8th.  It is possible I may complete the
work early; however, I need to budget for four full 12+ hour days.

The following resources will be shut down and UNAVAILABLE for some or all of
the four day window:

* All Compute Nodes.
* Auton Lab VPN
* GUARDDOG Cluster
* All TCWI Instances, including client facing instances.
* All /auton and /neill space.
* CVS and Subversion
* Autonlab.org mailing lists.
* Autonlab.org Jabber / XMPP services.
* All other services such as Bugzillas, etc.
* A minimalist site will temporarily replace the main www.autonlab.org
   website.

During this four day window, I will be leveraging the SCS Facilities imposed
shutdown to augment our computing infrastructure.  The work completed during
this time will provide:

* High availability improvements to /auton network storage.

* Improved network attached storage speeds.

* Improved scalability of virtualized server infrastructure.

* Improved recovery options when compute nodes crash.

* Streamlining and consolidation of firewall rules.

* Relocation of the following equipment into the rack:
   * LOW1 512 GB Compute node.
   * GUARDDOG Cluster (5 servers)
   * 3 Rackmount 240 volt 3000 VA UPS Units.

* Balanced and redistributed power connections.

* Server locations in racks will be moved as necessary.

* Upgrades of end-of-life operating system instances running CentOS 4 or
   Ubuntu 8.04.

Auton Lab Desktops will be functional, but they will not have access to
/auton space, Auton VPN or any of the usual resources listed under
"UNAVAILABLE", above.  This mail will be followed by a targeted mail
addressing the issues relating to Auton Staff Deskops.

Please notify me ASAP if you have any deadline situations which require
any of our central computing resources during Jan 4-8.  I will do my best
to provide fallback resources in the case of important deadlines which
cannot be avoided.


-- 
Michael J. Baysek
Systems Analyst
Carnegie Mellon University / Auton Lab
412-268-8939 - mjbaysek at cs.cmu.edu
http://www.autonlab.org