java - ElasticSearch Multiple Data Directories, choose where to place the index -


i'm running elasticsearch on multiple servers. servers equal, have 2 disks: 1 ssd, , 1 hdd. needless say, ssd faster yet smaller.

i know can set multiple data directories in es added paths elasticsearch.yml. but, default, (from i've found) es automatically chooses data directory take based on percentage of disk space available.

some indices more important me others, newer ones (those queried lot) need on ssd, , queried less can on hdd.

what need done - if possible @ time?
add index path?

this not possible in elasticsearch. yes, can specify multiple data paths, cannot "assigned" indices.

at moment, es stripe data on file level data paths, means shards spread on paths:

path directory store index data allocated node.

path.data: /path/to/data

can optionally include more 1 location, causing data striped across locations (a la raid 0) on file level, favouring locations free space on creation. example:

path.data: /path/to/data1,/path/to/data2

in 2.0 on other hand this change still not able have index on data path , index on data path.

at moment, solution use case have ssds on "hot" nodes , hdds on "cold" nodes, meaning 1 es node ssd , 1 es node hdd: https://www.elastic.co/guide/en/elasticsearch/guide/current/retiring-data.html#migrate-indices


Comments

Popular posts from this blog

toolbar - How to add link to user registration inside toobar in admin joomla 3 custom component -

linux - disk space limitation when creating war file -

How to provide Authorization & Authentication using Asp.net, C#? -