Constant Disk IO

[AdvancedStyle]

I'm running:

• Parity version: Parity/v1.9.0-unstable-52e2adc-20171019/x86_64-linux-gnu/rustc1.21.0 + PR Ethstore optimizations #6827
• Operating system: Ubuntu 16.04.3 LTS
• Hardware: RAM 10GB + SSD
• And installed: from source

Actual Behavior
After completely resyncing the chain, and waiting for a full db-compact the disk IO is still high, and there is a constant stream of data being written/read from disk. Even while waiting for new blocks to come in there is disk activity.

Maybe there is just some config I'm missing that will put an data to all this disk IO?

Current params:
parity --cache-size 4096 --allow-ips public --jsonrpc-apis web3,eth,net,parity,traces,rpc,personal --geth

Example output from iostat -x 1:

avg-cpu: %user %nice %system %iowait %steal %idle
27.57 0.00 3.76 4.51 0.25 63.91

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 0.00 977.00 0.00 7192.00 0.00 14.72 0.15 0.15 0.15 0.00 0.16 15.20

avg-cpu: %user %nice %system %iowait %steal %idle
22.90 0.00 7.12 5.34 0.00 64.63

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 20.00 803.00 43.00 5484.00 288.00 13.65 0.35 0.42 0.34 1.86 0.35 29.60

avg-cpu: %user %nice %system %iowait %steal %idle
30.53 0.00 15.52 3.56 0.00 50.38

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 48.00 819.00 64.00 6380.00 488.00 15.56 0.28 0.32 0.27 0.94 0.23 20.40

avg-cpu: %user %nice %system %iowait %steal %idle
25.50 0.00 5.00 0.00 0.00 69.50

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 23.00 7.00 20.00 64.00 196.00 19.26 0.01 0.44 0.00 0.60 0.44 1.20

avg-cpu: %user %nice %system %iowait %steal %idle
19.10 0.00 0.75 20.60 0.00 59.55

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 3.00 581.00 6.00 3648.00 1944.00 19.05 1.22 0.29 0.27 2.67 0.93 54.40

avg-cpu: %user %nice %system %iowait %steal %idle
34.10 0.00 9.41 22.90 0.00 33.59

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 68.00 1103.00 233.00 7584.00 3372.00 16.40 2.80 2.19 1.78 4.14 0.59 79.20

avg-cpu: %user %nice %system %iowait %steal %idle
37.72 0.00 15.95 6.08 0.00 40.25

Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util
vda 0.00 73.00 884.00 175.00 6448.00 1056.00 14.17 0.33 0.31 0.22 0.78 0.24 25.20

..

Expected Behavior
Cache DB data to memory and not constantly read/write from disk.

[AdvancedStyle]

I've made some adjustments to the params and seems to have totally solved the disk IO issue:

parity -- min-peers 10 --max-peers 20 --cache-size 6096 --allow-ips public --jsonrpc-apis web3,eth,net,parity,traces,rpc,personal --geth