CODEWITHSUNDEEP
amazon-web-servicesneo4j
drew moore
drew moore

Reputation: 32680

Neo4j failing to start in AWS instance

I've got an AWS instance (t1.micro in us-west availability zone) running Ubuntu 14.04, with neo4j-enterprise 2.2.3 installed.

I can start the service with an empty database without any problem, but when I try to start it with a database with a small amount (~200mb) of data (by modifying the datastore directory in conf/neo4j-server.properties), I get 

Failed to start within 120 seconds. 
Neo4j Server may have failed to start, please check the logs.

What's frustrating is that, when I look in /data/log/console.log, all I get is: 

2015-07-20 20:33:43.769+0000 INFO  [API] Setting startup timeout to: 120000ms based on 120000
2015-07-20 20:33:44.527+0000 INFO  [API] Successfully shutdown Neo4j Server.

and in neo4j.0.0.log: 

Jul 20, 2015 5:26:05 PM com.sun.jersey.server.impl.application.WebApplicationImpl _initiate
INFO: Initiating Jersey application, version 'Jersey: 1.19 02/11/2015 03:25 AM'

... which is to say, nothing helpful.

In /dbDirectory/messages.log, I have

2015-07-20 17:40:37.926+0000 INFO  [org.neo4j]: Thread[main,5,main] Starting BatchInserter(EmbeddedBatchInserter[/var/lib/neo4j/data/civ_sm.db])
2015-07-20 17:40:37.946+0000 INFO  [org.neo4j]: Creating new db @ /var/lib/neo4j/data/my.db/neostore
2015-07-20 17:40:38.251+0000 INFO  [org.neo4j]: No lucene scan store index found, this might just be first use. Preparing to rebuild.
2015-07-20 17:40:38.278+0000 INFO  [org.neo4j]: No lucene scan store index found, this might just be first use. Preparing to rebuild.
2015-07-20 17:40:38.403+0000 INFO  [org.neo4j]: Rebuilding lucene scan store, this may take a while
2015-07-20 17:40:38.404+0000 INFO  [org.neo4j]: Lucene scan store rebuilt (roughly -1 nodes)
2015-07-20 17:41:04.646+0000 INFO  [org.neo4j]: Missing counts store, rebuilding it.
2015-07-20 17:41:07.377+0000 INFO  [org.neo4j]: About to rotate counts store at transaction 1 to [/var/lib/neo4j/data/my.db/neostore.counts.db.b], from [/var/lib/neo4j/data/my.db/neostore.counts.db.a].
2015-07-20 17:41:07.382+0000 INFO  [org.neo4j]: Successfully rotated counts store at transaction 1 to [/var/lib/neo4j/data/my.db/neostore.counts.db.b], from [/var/lib/neo4j/data/my.db/neostore.counts.db.a].
2015-07-20 17:41:07.606+0000 INFO  [org.neo4j]: Thread[main,5,main] Clean shutdown on BatchInserter(EmbeddedBatchInserter[/var/lib/neo4j/data/my.db])

(I wrote a small program that generated the data in this DB using BatchInserter, and the contents of this log all seem to be related to the creation process. There is nothing pertaining to the failed start.)

Attempting to isolate the problem, I downloaded the database to my local machine (running Fedora and the same (2.2.3) neo4j version), and it started without issue.

Two other notes - I've set max-open files to 40k (though that shouldn't be an issue given the limited size of the db), and the failure occurs in ~2 seconds

Upvotes: 3

Views: 258

Answers (1)

drew moore
drew moore

Reputation: 32680

Purely for the sake of marking this question as resolved - @MichaelHunger nailed it in his comment: This was a perms issue - I had mistakenly chown'ed /.../neo4j/data/ appropriately but not /.../neo4j/data/theSpecificDbDirectory/

I will say, as much as I love neo4j (which is a lot), it's more frequent than I'd like that I wish its log messages were a little more helpful, and this was one of those cases...

Upvotes: 1

Related Questions