I heard Local / NVME SSD were used on Elastic Cloud on GCP, I wonder if the local SSDs configured for Elastic Cloud on GCP utilize the LVM to merge individual disks into single logical partition or RAID?
I don't seems to find any info for that on the configuration page.
One issue we faced during prototyping off-the-cloud is that with multiple local SSD disks attached, the distribution of the data is not even amongst the attached disks, causing the node to hit disk usage watermark when only one of the disks uses excessive amount.
One example below:
Filesystem Size Used Avail Use% Mounted on
devtmpfs 32G 0 32G 0% /dev
tmpfs 32G 340K 32G 1% /dev/shm
tmpfs 32G 540K 32G 1% /run
tmpfs 32G 0 32G 0% /sys/fs/cgroup
/dev/sda2 50G 6.4G 44G 13% /
/dev/sda1 200M 12M 189M 6% /boot/efi
/dev/nvme0n1 369G 270G 100G 74% /data/fs1
/dev/nvme0n2 369G 240G 130G 66% /data/fs2
/dev/nvme0n3 369G 267G 103G 73% /data/fs3
/dev/nvme0n4 369G 343G 27G 93% /data/fs4 <========
/dev/nvme0n5 369G 287G 83G 78% /data/fs5
/dev/nvme0n6 369G 242G 128G 66% /data/fs6
/dev/nvme0n7 369G 248G 122G 68% /data/fs7
/dev/nvme0n8 369G 247G 123G 67% /data/fs8
I read about cluster.routing.allocation.balance.disk_usage setting and wonder if it would help in the case?
Thoughts?