...
- Scott Prater
- Setting up cluster following Frank's wiki page
- Should be a couple more hours
- Will be getting testsuite from Frank
- Setting up cluster following Frank's wiki page
- Osman Din
- Working on ticket for hiding update ability on read-only federation
- Mike Durbin
- Wrapping up version ticket
- There are many endpoints that are fedora-specific
- Will be stopping when fcr:content works with versioning
- Mike Daines
- Finished some unit tests
- Will then be starting fcr:export bug
- Greg Jansen
- Working on refactoring config
- Frank Asseg
- out
- Eric James
- Submitted PR for type hierarchies
- David Wilcox
- First day on the job
- Ben Pennell
- Still working unit tests, auth-roles-commons
- Adam Soroka
- Working OSGi indexer,
- Will begin producing ontology mappings for the JCR -> Fedora -> LDP sequence
- Will lay some comments on Kai transaction-related PR, when he provides a new PR from his fork
Tuesday February 18
- Scott Prater
- Cluster is up and running
- Reviewing JGroups and ISPN configuration/documentation
- Will start running Frank's tests
- Pretty far along with puppet module
- Cluster is up and running
- Osman Din
- out
- Mike Durbin
- Updating unit tests for versioning
- Will have PR soon
- Will start transaction tickets next
- Updating unit tests for versioning
- Mike Daines
- Addressed PR code comments
- Have been working on /fcr:export
- Issues with Jersey
- Greg Jansen
- Working on consolidated configuration
- Also addressed auth-war
- Action: Andrew to create ticket for auth additional config
- Starting puppet clustering and load-balancer
- Working on consolidated configuration
- Frank Asseg
- out
- Eric James
- Finishing ticket regarding rdf hierarchy types
- Ben Pennell
- Submitted PR for unit tests
- Will be looking through backlog for more tickets
- Adam Soroka
- Switching from OSGi and to JCR/LDP/Fedora mapping strawman
- After sprint meeting: Scott Prater/Greg Jansen
- Discussion: how does clustering work? Infinspan documentation seems to suggest that a shared persistent file store is the best way to go: https://docs.jboss.org/author/display/MODE/Clustering#Clustering-Replicated
- Scott notes that nodes seem to be volatile, come and go, are tied to processes: Greg agrees (cluster is mostly about managing availability and load balancing)
- Scott describes Tomcat shutdown problem, the nasty side-effects it has on stopping, restarting nodes in a cluster: https://www.pivotaltracker.com/story/show/65988656
- Next steps: Scott: finish up puppet-fcrepo module, so Greg can use it; also, begin to duplicate Frank's tests
Greg: finish installing and configuring Puppet on his cluster, get cluster up using puppet-fcrepo module
...