LexEVS Meeting Minutes - 2014.10.15

De Coronado, Sherri    

Safran, Tracy

Kuntipuram, Kumar

Scott to work with Gilberto to determine what is needed for OWL 2 requirements

on hold

on hold

on hold

on hold

done

Scott: CTS2 deployment, SNOMED/ ICD10CM issues, Issues support

• • Support - 3 external users.  One from ASU, one from Spain, and one other.   Indications that we need to have improved documentation.   
    • Eric - Scott to follow up (from Gilberto) - looking to set up a federated set of CTS2 notes for German universities
    • Some of the issues - mismatches of GIT content being used; support of non-tested versions of tech stack.  
    • May want to consider "bootcamp" like education - was we did back during the caBIG days 

Cory: CTS2 issue and AMIA Poster

• • AMIA poster sent to authors
  • CTS2 search issues
  • JIRA issue

Craig: Project Plan, Project Startup, Poster

• Plan on sending plan out this week
• Initial product dev documentation section of wiki created.  Will add documents as required.
  
  

Loads at NCI are acting differently than our loads here at Mayo.

Plan on reload transitive table (after DB restart).  Then look at SQL settings/variables in configuration to determine if there are differences. 

ICD10 CM - comments from Joanne and Stephanie.   It was decided that LexEVS is now doing what is needed, but Stephanie needs to work out some content issues.  

Cory and Scott investigated and determined that index is missing in the DB - need to have indexes built on QA to reflect the database on QA.  Rob working on getting the indexes on Dev - going to bounce the DB and then reload.  If successful, then move the indexes to QA.  

Stage currently not avaible to outside users for testing.  Want to move Stage to the same segment as Production. 

Separate STAGE server in the production network segment.   (1 production blade and 1 stage blade)

Would prefer to have a fail-over blade. 

This is a security policy in that the servers that are on an external segment need to be APP scanned.   There cannot be a non-app scan server on the network with external accessible servers (PROD).

There are 2 blades provisioned in the production environment (Blade A is primary and Blade B is fail over)

There is a need to isolate tiers from other tiers.

Suggested to use the second Blade (with option to use the blade if primary blade goes down).

Need to consider I/O contention, lucene indexes, paths, etc, if 2 instances.   There is sufficient DB storage.  100GB was planned in the arch document.  (2 instances (STAGE and PROD) on the same blade server)

There are currently 4/socket and 4 sockets per blade server.  

LexEVS was first to be deployed in this new environment.  

Blade A is running production (like Stage on lower tier).  Blade B is running differently than production. 

It was suggested to start with the 2 Blades in production as A - Production and B - Stage.

As procurement processes get better, then consider additional servers.  

Lucene on SAN-based (8 GB fiber, not internal drive) - would this work?  Would normally try this in QA.  

Collaborators are testing software outside the firewall. 

Diagrams to be created to visual the environment. 

Possible discussion on Wednesday PM or Thursday AM/PM.  

Thursday could work.  Need to come up with agenda/topics.