Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

High disk-utilisation with non workload

Hello,

I 'm running a 3 node petasan cluster with 17 NVME-OSDs inside. Each  node has 4x40GBit Intel NICs,20 Cores, 128GB. I 'm only using  RDB/iSCSI. The highest troughoutput I've ever seen was poor 300MB/s vs Windows 2019 MPIO.

The Dashboard shows 100% disk-utilisation the whole time, no matter the value of Cluster troughoutput. The internal benchmark shows the same thing.


Disk tps rkB/s wkB/s areq-sz aqu-sz await svctm %util
nvme1n1 578 2230 1745 6 1 0 1 99
nvme0n1 1143 24223 2587 23 1 0 0 99
nvme2n1 345 1353 1169 7 1 0 2 99
nvme5n1 608 2467 357 4 1 0 1 99
nvme4n1 644 2565 1691 6 1 0 1 99
nvme6n1 455 1806 882 5 1 0 2 99
Cluster IOPS :
  • Write  23561
  • Read   53515

Node Memory Util% CPU Util% Network Util% Disks Util% Actions
Avg Max Avg Max Journals OSDs
Avg Max Avg Max
psn03 29 25 31 1 2 0 0 100 100

What can I do?

Kind regards Hartmut

 

This is kernel issue

https://github.com/sysstat/sysstat/issues/187