See our general page on upgrading clusters for current list of considerations regarding upgrading Matrix.

Ticket number for procurement-related email conversations

See also

• Matrix end-user documentation

Ticket numbers for email conversations

• INC000001223799

  • Scheraga - Matrix testing Phase 2 (//all// researchers confirming their applications)

  • Created 10/15/2014, Wednesday)

• INC000001120391
  • Planning the implementation, and for more generic issues with the new Matrix and related processes.
  • Created
  INC000001120391 (new as of
  • 6/6/14, Friday)

REMINDER to ChemIT staff: Just put this one of these number in the Subject line, and send message as ChemIT and to ChemIT (BCC if fine). That's it!

N.B. For Deprecated (now old) tracking numbers

• Tracking number that was used for the spec'ing and ordering phase

...

• : INC000001055288
• Tracking number that was used for the Phase 1 software testing, with 7 initial software testers: INC000001202826

Resources

• https://cornell.app.box.com/files/0/f/2132768200/Matrix

Decisions required before purchase

Note: Order must be $5,000 or larger to not pay "overhead". Therefore, buy parts, especially small ones, together. (And a single purchase is better than multiple, $5,000+ purchases.)

• ChemIT's proposed specifications and prices for Matrix upgrade.

Decisions made

Compute nodes

NOTE: Must finalize quantity (and final, final specs) after all other decisions are made and priced.

4/14/14. Czarek: It looks like the slowest cpu E5-2620v2 2.1GHz has the best price performance ratio but anyway I would not buy the slowest cpu. Right now in Gdansk we are buying 10 servers and we decided to go for 10-core cpus E5-2670v2 2.5Ghz (20 cores per node). As in matrix in Gdansk we have only slow interconnect between nodes and some programs can run efficiently only on single node so more cores per node gives for such program advantage. Other programs both in Gdansk and on matrix just need the highest possible total performance and exchange very little data between nodes so than number of cores is not important. What about space restriction ? Is it better to buy smaller number of faster nodes ?

Information ChemIT needs to keep track up, for reference

See sub-pages

Decision 1: Storage of users' data

Where should users' data be located?

• Current answer: Hard drive RAID array in the head node.
• Choices identified by ChemIT, to work through with Czarek:

Decision 2: Backups

How should Matrix be backed up?

Answer is DEPENDENT on Decision 1.

• Must include capacity to restore OS, applications and configurations, and users' data.
• Current answer: Cornell CIT's EZ-Backup.
  • Details: See "Scheraga - Cluster backup choices" for data informing this original decision (created following the Fall 2013 Matrix failure)
• Choices proposed by ChemIT, to work through with Czarek:
  • TBA: Backup may be included in options considered for Decision 1.
    • If still required, ChemIT still recommends EZ-Backup.

Decision 3: Head node specifications

The head node specifications is DEPENDENT on Decision 1 (and maybe Decision 2).

• Size decision
  • 1U is cheapest, and fine if don't need to hold lots of drives.
  • 2U is more expensive, but can hold more drives.

Considerations for upgrade (info-gathering phase)

...

...

Implementation decisions (for after purchase decisions!)

Storage-related

For each question:

• Current answer:
• Proposed by ChemIT:
• Proposed by Czarek:

Where should applications be installed and configured?

Where is "temp" (not backed up) data located?

Who gets "root" access?