doubt with Neo4j HA

[D.Radhakrishnan]

Hi,

saw this exception in the slave's data/graph.db/messages.log--

2012-08-18 07:17:06,469 INFO  [neo4j]: Started - database is now available

2012-08-18 07:17:06,469 INFO  [neo4j]: GC Monitor started.

2012-08-18 07:17:06,471 INFO  [neo4j]: Stopping - database is now unavailable

2012-08-18 07:17:06,472 INFO  [neo4j.txmanager]: TM shutting down

2012-08-18 07:17:06,473 INFO  [neo4j.xafactory]: Closed log /home/pg/neo4j-enterprise-1.7.1/neo4j-enterprise-1.8.M06/data/graph.db/temp-copy/index/lucene.log

2012-08-18 07:17:06,475 INFO  [neo4j.xafactory]: Closed log /home/pg/neo4j-enterprise-1.7.1/neo4j-enterprise-1.8.M06/data/graph.db/temp-copy/nioneo_logical.log

2012-08-18 07:17:06,481 INFO  [neo4j.datasource]: NeoStore closed

2012-08-18 07:17:06,481 DEBUG [neo4j.diagnostics]: --- SHUTDOWN diagnostics START ---

2012-08-18 07:17:06,481 DEBUG [neo4j.diagnostics]: --- SHUTDOWN diagnostics END ---

This exception is there in neo4j.0.0.log,

Aug 18, 2012 7:17:09 AM org.neo4j.server.logging.Logger log

INFO: Server started on [http://0.0.0.0:7474/]

Aug 18, 2012 7:17:52 AM java.util.prefs.FileSystemPreferences$2 run

WARNING: Couldn't create user preferences directory. User preferences are unusable.

Aug 18, 2012 7:17:52 AM java.util.prefs.FileSystemPreferences$2 run

WARNING: java.io.IOException: No such file or directory

Aug 18, 2012 7:18:23 AM java.util.prefs.FileSystemPreferences checkLockFile0ErrorCode

WARNING: Could not lock User prefs. Unix error code 2

I have attached the messages.log and ne04j.0.0.log .Can you pls tell me what went wrong. am stuck at configuring a neo4j  HA since last week :( 

[Peter Neubauer]

Hi there,
It seems there is a file locking issue from the logs. Could you check that the file permissions give the database process user write access to all files in the database directory?

/peter

Send from mobile.

--
 
 

[D.Radhakrishnan]

Hi peter, 

the issue is resolved now.

Now i have some doubts with loadbalancer, REST and HighlyAvailableGraphDB.

When we have HighlyAvailableGraphDatabase, we give the following 

ha.server_id

ha.server

ha.coordinators

suppose, i have a master + 2 slaves, what value should i give for ha.server_id and ha.server.

ha.coordinates contain the IPs of the zookeeper cloud. (am clear with this one ) 

The moment this line is getting called   graphDatabaseService = new HighlyAvailableGraphDatabase( "data/graph.db", config ); the entire DB data that is there in the master server is replicated to the client. Is that expected? if it is, can we stop the entire copying ?

How will we use the HA_PROXY load balancer in case of HighAvailableGraphDataBase or HA_PROXY is meant only for REST API.

finally, how to load balance with HighAvailableGraphDataBase if HA_PROXY is meant for REST.

Thanks.

[Peter Neubauer]

Hi there,
I am copying the explanations in the neo4j.properties here ... let me
know if they do not make sense and we can adjust things.


# Uncomment and specify these lines for running Neo4j in High Availability mode.
# ha.server_id is a unique integer for each instance of the Neo4j
database in the cluster.
# (as opposed to the coordinator instance IDs)
# example: ha.server_id=1
ha.server_id=1

# ha.coordinators is a comma-separated list (without spaces) of the
host:port of where to
# find one or more of the Neo4j coordinator servers.
# Avoid localhost due to IP resolution issues on some systems.
# example: ha.coordinators=localhost:2181,1.2.3.4:4321
ha.coordinators=localhost:2181,localhost:2182,localhost:2183

# You can also, optionally, configure the ha.cluster_name. This is the
name of the cluster this
# instance is supposed to join. Accepted characters are alphabetical,
numerical, dot and dash.
# This configuration is useful if you have multiple Neo4j HA clusters
managed by the same
# Coordinator cluster.
# Example: ha.cluster_name = my.neo4j.ha.cluster
ha.cluster_name=local.jvm.only.cluster

# IP and port for this instance to bind to to communicate data with the
# other neo4j instances in the cluster. This is broadcasted to the other
# cluster members, so different members can have different communication ports.
# Optional if the members are on different machines so the IP is
different for every member.
ha.server = localhost:6001

Cheers,

/peter neubauer

G: neubauer.peter
S: peter.neubauer
P: +46 704 106975
L: http://www.linkedin.com/in/neubauer
T: @peterneubauer

Wanna learn something new? Come to @graphconnect.

> --
>
>