Hi!
We found out that one of our nodes in cluster uses 100% of nvme drive, that is weird as outputs of any disk analiser shows that there are no excesive usage of disk drive.
![1673865778323.png]( "1673865778323.png")
One process uses 97% of IO while total disk wirte/read is way below declared by samsung nvme specification.
![1673865837994.png]( "1673865837994.png")
zabbix shows 100% usage from this day and in the night
![1673865874920.png]( "1673865874920.png")
This causes IO delay and I can't find the reason
Package verions below:
Read more
We found out that one of our nodes in cluster uses 100% of nvme drive, that is weird as outputs of any disk analiser shows that there are no excesive usage of disk drive.
One process uses 97% of IO while total disk wirte/read is way below declared by samsung nvme specification.
zabbix shows 100% usage from this day and in the night
This causes IO delay and I can't find the reason
Package verions below:
Code:
proxmox-ve: 7.0-2 (running kernel...Read more