Control over writing data to a specific drive

We are planning to use single node Elasticsearch cluster. The node may contain multiple locations for writing indexing data.

I would like to know, if it is possible to write the data to a specific indexing location only?

For example, if there are 2 drives added in namely C: & D: in a single node. Does elasticsearch provides a control to write the data to only C: drive?

Let me know if you've any idea how to achieve this.

This functionality is deprecated and currently being removed, you probably shouldn't plan to use it:

Use one node per path.

1 Like

Thanks for letting me it is going to depricate in the next release.

We are planning to host an elasticsearch cluster on a single virtual/physical machine. OS will be windows.

Can you please let me know, if we can deploy 2 nodes on a single cluster?

Please share if there are any documents around.

Use one node per path.

Hi David,

If we can use only one drive then in case of disk full scenario, how can we move to another drive? What about the existing data?


Use one node per drive. If you want to move data off a drive, you can use shard allocation filtering.

Is there any mockups or documents available for "one node on path" design? Would like to know what all things one should consider during planning phase?

I'm not sure what sort of documentation you might need. You configure one node for each data path, it really is that simple.

I thought I'd mentioned this above but it seems not: you can combine multiple drives together into a single data path using RAID-0 or similar. On Linux there are options like LVM too, not sure if there's a Windows equivalent to that. That's the simplest way to scale a node up as your needs grow.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.