Hey Aniket,

Yeah we usually discuss on the tickets just to keep it in one place but
either is totally fine.

1. Actually that wasn't quite what I was proposing. What I am saying is
that there are three cases (a) metadata data missing in all dirs, (b)
metadata missing in some dirs, (c) metadata inconsistent between dirs. In
the case of (a) we should generate an id, in the case of (b) we should fill
in the missing data (this would be the case where a drive is destroyed and
replaced), and in the case of (c) someone has done something sketchy and we
should just error out. Let me know if you think that makes sense. An
alternative approach would be to designate a special place to keep this
kind of metadata but the question is always what happens in the case of
drive failure with multiple independently mounted drives.

2. Yup. We have a metadata api that does this.

On Wed, Oct 2, 2013 at 9:50 AM, Aniket Bhatnagar <[EMAIL PROTECTED]
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB