restoring a corrupted DB

[j2d3]

Hi, I have a db that somehow got corrupted. I am trying to start neo4j up, but it tries for a while and says ....BAD

What can I do to try to restore the data / get neo to start up again with that data?

(using Neo4j - Graph Database Kernel 1.8.M07)

[John Durkin]

So, been doing more analysis of what happened...

suddenly, the database started choking and spitting out errors as fast as it could write to messages.log (sample errors pasted here)

http://pastebin.com/dve3Exa8

This continued right up until the partition was filled.

We've never seen this sort of error before in our logs and we have not figured out what happened to trigger it. (This means we're kinda scared we will cause it to happen again).

Any ideas?

This is the enterprise version (Neo4j - Graph Database Kernel 1.8.M07)



Best,
JD

--
 
 

[John Durkin]

one more related question:

is the location of messages.log configurable?

[Peter Neubauer]

John,
let me talk to the team and get back to you.

Cheers,

/peter neubauer

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

Neo4j 1.8 GA - http://www.dzone.com/links/neo4j_18_release_fluent_graph_literacy.html

> --
>
>

[Mattias Persson]

This is a bug which has been fixed in 1.8 GA. So just download that and run and you'll be fine. It's not corruption btw.

2012/10/15 Peter Neubauer <peter.n...@neotechnology.com>

--

--
Mattias Persson, [mat...@neotechnology.com]
Hacker, Neo Technology
www.neotechnology.com

[j2d3]

Hi Mattias!

Thanks for the reply. So - I'm upgrading everything to 1.8 GA. What is the bug - the massive crazy error logging? Or the fact the db won't start up - I mean - If I upgrade to 1.8 GA, are you saying that the data I have might actually be okay, not corrupted, and actually start up properly?

I'm going to try it!

JD

[Mattias Persson]

Yes, that's what I'm saying and I'm quite sure it'll work out :) The exception spamming I gather is due to each write request failing and generating this error. When you shut down and tried to recover it would run into the same issue, namely a bug with rotating the lucene log. The bug caused the internal state (nothing to do with what's on disk) in the lucene data source used by neo4j.

2012/10/15 j2d3 <john....@gmail.com>