What I'd like to see added to the SolrCloud wiki page:
- The wiki page states that you can send your request to any server, but
what if that server goes down? Doesn't there need to be an aliased IP
address pointing to an active server? Or, is there client side support
like MongoDB replica sets where the client gets a seed list of nodes,
then gets a complete list, and the client can send requests to alternate
nodes if one doesn't respond.
- Are all the possible cloud configurations described or implied by the
examples on the wiki page? Is it possible to have only 1 shard and 1
replica? Is it easy to then split the single shard into more as you add
new nodes? How much of organizing the shards/replicas is automated and
what should be manually configured for optimal performance for my situation?
Those are the areas which are still unclear to me after reading over the
wiki page. I'm sure I'll have more questions as I move forward to
integrating the SolrCloud with our searching.
Shaddy
On 10/12/2012 11:14 AM, Mark Miller wrote:
On 10/12/2012 01:42 PM, Shaddy Zeineddine wrote:
Hello,
I have some questions about the SolrCloud.
Can I take full advtange of the Cloud with the PECL Solr client? It
was last updated for Solr 3.1 http://pecl.php.net/package/solr
I don't know for sure, I don't know that client. If it's HTTP based, it
probably still works, but somehow else might no better. The answer
probably involves some testing.
Is Jetty the recommended servlet for the Cloud?
It's what I would recommend.
The documentation about configuring, optimizing, and accessing the
Solr cloud is very sparse. How can I get more information about these
topics?
What would you like to see added to the SolrCloud wiki page?
- Mark
Shaddy Zeineddine, PHP Developer
szeinedd...@breakmedia.com | 310.360.4141 Ext. 297
Explore: http://acumen.breakmedia.com/ - Trends, Insights and ideas about Men
