On 9/20/2018 8:44 AM, Schaum Mallik wrote:
Thank you for your detailed responses. I am still kind of confused though.
Just to give you some more insight. When I first created the cloud I
created to collections and the ‘-d’ option pointed to the directory where
the config for the collection was stored which was in
/opt/solr/server/solr/configsets/articles/conf
so as you are saying if I move the replica data to another location and
restart the node and most probably solr might complain then what happens
next. This solr node was working fine a week back and the indexes were
upgraded when they were moved from solr 6 to solr 7
I really think that those cores that won't initialize are remnants from
a non-cloud install and probably have *never* worked. For some reason,
you didn't notice these errors until now.
If you take a non-cloud Solr instance that has indexes and add ZK_HOST
to make it a cloud instance, the cores that existed before cloud mode
will NOT work.
I just completed a test where I did that exact sequence of converting a
node from non-cloud with existing indexes to cloud, and ran into the
exact same errors you're seeing. I can relay exact details of the test
if required.
I think that moving the non-cloud cores out so Solr doesn't try to load
them will likely fix this issue.
Thanks,
Shawn
