Right, that makes sense and it's really what I just said in my reply to Ryan's
email. But currently only data/index* is being replicated, unless I'm
misreading this part of snapshooter:
# take a snapshot using hard links into temporary location
# then move it into place atomically
if [[ "${OS}" == "SunOS" || "${OS}" == "Darwin" ]]
then
orig_dir=$(pwd)
mkdir ${temp}
cd ${data_dir}/index # here here
find . -print|cpio -pdlmu ${temp} 1>/dev/null 2>&1
cd ${orig_dir}
else
cp -lr ${data_dir}/index ${temp} # here here
fi
mv ${temp} ${name}
Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch
----- Original Message ----
From: Yonik Seeley <[EMAIL PROTECTED]>
To: solr-user@lucene.apache.org
Sent: Friday, December 7, 2007 1:16:19 PM
Subject: Re: spellchecker and multi-core index replication
On Dec 7, 2007 12:53 PM, Otis Gospodnetic <[EMAIL PROTECTED]>
wrote:
> Hi,
>
> I wrote the original spellchecker request handler a while back and a
lot of people seem to be using the spellchecker stuff, fixing and
improving it! :)
> But recently I realized I never wrote anything that takes care of the
spellchecker index replication. Looking at snapshooter, it's
completely unaware of the spellchecker index -- or *any* other index in there
-- it explicitly looks for data/index.
I originally came up with data/index with the idea of taking snapshots
of the whole data directory...
/solr/data
/solr/data/index (the main index dir)
/solr/data/foo
/solr/snapshots (snapshots of the complete data directory)
-Yonik
