An ancient programmer that has not really coded for about 40 years, brought up on Commodore Basic and 6502 assembler.

Now trying to learn enough JAVA and Elasticsearch to make a little "niche" research tool version of a search engine.

Budget is virtually nothing, equipment are six twin-cpu DL360-Ver 9 with 128GB ram. Three for the main cluster, three more when the data starts to grow. The idea is all these nodes will be dedicated machines and run Master+Data+Client.

Then a DL560 which will be INGEST only on a SHARED machine so it can pre-process new data - and then join the cluster as and when it needs to ingest,

Trying to learn about how best to allocate resources, anticipate index size (documents are going to be 100 million+ but only about 5K each, plain text).

Trying to learn from Google and other posts, but it gets challenging when half of it contradicts the other half. Still, wouldn't be fun if it was too easy.

It's a steep learning curve though

(Guy) Mark