Multiple data paths, Force Merge API and Disk Size

I have few questions relating to multiple data paths

  1. Assume I have multiple data paths like below /mnt/first,/mnt/second

    Does Force Merge API work ( max_num_segments = 1 ) in this scenario (where index data has been written on multiple disks) ?

  2. My application data will max use 15 TB data. So If I plan use the AWS EBS st1 of 16TB (that's the max it can go). Will there be any performance issues if I use a single disk of 16TB or should I assign multiple volumes (multiple data paths, say 2 x 8TB, or 4x4TB)? Is there any recommended best practice here # of paths one should use?


Regarding your first point, forceMerge is segments per shard. Since shards do not span data paths (not anymore--since 2.x, IIRC), a forceMerge will not cause issues in the way you are suggesting.

Thanks Aaron.

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