Just got off a call with Mongo expert... Seems it's not good for large systems that need to be fast and scale...

Really? I thought it was all about scaling? (I know bare minimum about it)

@Hobo42 so did I... Thought it was basically the hadoop solution... Seems like it actually isnt and when it we do it it gone be huge pain in the ***

Don't know who that expert is, but it's a vague response..

Using MongoDb for a little blog, is like building a villa to store your peanuts.

Whenever it comes to large datasets - it's a matter of how you store, use and maintain it. Usually you would store everything raw, then run queries on that to generate reports. Live queries will run slow on huge datasets regardless of technology. Aggregated reports (documents) that are appended as data changes is what you will query upon to your end user. Typically, specialized reports suited for the task at hand.

Not knowing your data, but you can't expect instant result on something that can't be stored in memory and require different documents/collections to join.

And.. Don't use a single server instance for everything. Your raw storage should be kept alone. Optimize your report servers for read.

@viking8 will that's the opened, don't use single server instance, it seems it's cluster set up needs more servers but isn't as good as Hadoop either, not efficient out able to distribute the data optimally by itself

MySQL (Maria DB) for life...