What advantage do you see in TLOG and/or PULL replicas? The TLOG and
PULL replica types are for some pretty specific use cases,
particularly high-indexing-throughput cases where you can't afford to
index each doc on every node in your cluster. If you can afford the
CPU cycles to index on every
Hi Erick/Robi,
Thanks for your replies! one more question, if I go with solrcloud and
having PULL/TLOG replication mixed in the cluster, by the documentation,
this would have Master/Slave mode benefit as well, does that seem feasible
to you? are there any projects adopting this new feature in
On 12/14/2017 2:27 PM, Drooy Drooy wrote:
> We have an in-house project running in Solr 4.7 with Master/Slave mode for
> a few years, what is it going to take to upgrade it to SolrCloud with
> TLOG/PULL replica mode ?
>
> I read the upgrade guides, none of them talking about the jump from 4.x to
>
Completely agree with Robert. I'd also add that you should _not_ copy
your configs from 4x. Start with the 7x configs and add any
customizations but don't change things like luceneMatchVersion and the
like.
If you simply _cannot_ reindex, take a look at
>From what I have read, you can only upgrade to the next major version number
>without using a tool to convert the indexes to the newer version. But that is
>still perilous due to deprications etc
So I think best advice out there is to spin up a new farm on 7.1 (especially
from 4.x), make a
