9000v - google cloud -24vCPU-60Gig-RAM - unstable

[Marway]

Hello All,

I have GCP EC instance, 24vCPU, N1, 60Gb RAM.
I was abble to run max 4 x 9000v (9.3.3 version) in stable manner.
If I run 6x 9000b, they either randomly reboots, or load to bash ... similar issues as other users on the forum, when they used weaker CPUs or didn`t have enough of memory.

I have assigned 4 vCPU & 8k RAM per 9000v Node.

When i check the CPU on GCP instance, I can see that it has :
24x
model name : Intel(R) Xeon(R) CPU @ 2.30GHz
cpu MHz : 2300.000

Only way how i can get stable solution (without reboots) is with 4x9000v ..and 4 only

May you share your experience ? is there any way how to run 6 + x 9000v in google cloud ?
Do you think i need more powerful CPU - N2 ?

P.S.
- i have tested out 4x 2vCPU; than 6x 2vCPU ... same results => unstability
- that is why i tested 4vCPU per 9000v, and tried 5x 9000v ... and solution was still unstable ..

Thank you

Marway

[Marway]

I have tested N2 CPUs ... I have now 32vCPU + 128 GB RAM.
I have issues even with 4 x 9000vs ...
getting crashes from
!
....[ 382.012789] NMI watchdog: BUG: soft lockup - CPU#1 stuck for 24s! [clis:8585]
[ 382.012936] NMI watchdog: BUG: soft lockup - CPU#0 stuck for 24s! [cfs:9247]
[ 382.013380] Modules linked in:
!
Not sure if this is related to Qemu or not.