RavenDB 3.5

Preview:

Citation preview

2ND RAVENDB CONFERENCEOren Einioren@ravendb.net

RAVENDB 3.5Oren Einioren@ravendb.net

STATS• More than 2,500 commits• Work started in Dec 2013• 20+ core dev team• Over 20 major features

FOCUS• Distributed work• Operations• Performance• High load

DISTRIBUTED WORK

RAFT IMPLEMENTATION• The cluster

• Reliable• Cluster wide operations• Dynamic leader

• Replication• Multi master• Clients will use leader as primary• No jitter on failover/recovery• Reduced conflicts

LOAD BALANCING SLA

I choose you the fastest

FILTERED REPLICATION

CAN YOU EXPLAIN THAT?

LARGE CLUSTER IMPROVEMENTS

OPERATIONS

ADMIN CONSOLE

GLOBAL CONFIGURATION• Simplify multi database configuration• Backup• Replication• Versioning

I/O INSIGHTS

I/O MONITORING

REQUEST TRACING / TIMING

DATA EXPLORATION

SCRIPTINGwith document_store.open_session() as session:

people = list(session.query(Person).where(name="Arava"))for person in people:

print person.name

 

SNMP MONITORING

zabbix.ravendb.net

VERSIONING

PERFORMANCE

THREAD POOL• Dedicated indexing threads• Dynamic priorities• Work stats• Smarter work distribution

FULL TEXT SEARCH• Optimized analyzer• Reduced allocations

QUERYING• Memory mapped indexes• Better query parser

• No exceptions!

MEMORY• Low memory handling• Lower memory usage

• Many concurrent indexes• Slow I/O systems

• Tracing auto tuning decisions

QUESTIONS?

Recommended