Hi Vinay,
1) No, it's not advisable to use the same cache for different types of
objects. The example from the article had a mistake. Fixed the example
in a way that every object of an unique type is stored in a specific
cache. Thanks for noting this!
2) Since replicated cache stores a copy of every object on every node
from the cluster there is no need to use affinity keys for objects of a
replicated cache.
3) Correct, the objects should be either collocated (using affinity key)
or stored in a replicated cache.
You can refer to org.apache.ignite.examples.datagrid.CacheQueryExample
to see how affinity collocation works in practice. Pay attention to how
caches are initialized, affinity keys are set up and used in the example
and how distributed SQL query works.
Regards,
Denis
On 1/9/2016 9:02 PM, vinshar wrote:
I went through documentation for enabling Affinity at below link
https://apacheignite.readme.io/docs/affinity-collocation
Following are my queries:-
1) Document has an example where same cache is used to hold different type
of cache objects. Is this advisable and is this necessary to achieve
collocation of objects?
2) Is there any case where it may make sense to use Affinity Key's in case
of a replicated cache?
3) For cross cache SQL Queries, Cached objects should be collocated. If this
is correct then how can i achieve this?
Regards,
Vinay Sharma
--
View this message in context:
http://apache-ignite-users.70518.x6.nabble.com/Cross-Cache-Affinity-tp2464.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.
