ver: Truenas 12.0-U2
Running as a vm under esxi 7.0u1.
My internet connectivity is symmetrical gigabit fiber.
I notice during high network activity, completely unrelated to the nas (ie running a speed test on a desktop), truenas cpu usage is up to 30% per esxtop. Running top within truenas at the same time shows no processes consuming any significant cpu.
The only truenas services running are NFS, SMB, Smart and SSH. The only common factor is the network activity and the fact that the nas is on the same subnet, portgroup and vswitch as the client running speed test. Other vm's set up similarly (nextcloud/ubuntu, cyberpower powerpanel/centos, freepbx/ubuntu) don't exhibit this behavior.
I have a few vlans set up for iot purposes. Speedtests performed on different vlans also seem to cause this excessive cpu.
On a whim, I tried to isolate truenas to a separate portgroup bound to the same vswitch. Config looks like this.
Previously, Truenas.nas was part of the LAN_port(2) portgroup. Repeating the speed tests on the network no longer causes high cpu in truenas.
I don't see any negatives from the updated config but am still puzzled¿? Why this is happening in the first place? Network traffic is still being generated on the same vswitch and everything remains fully accessible.
Running as a vm under esxi 7.0u1.
My internet connectivity is symmetrical gigabit fiber.
I notice during high network activity, completely unrelated to the nas (ie running a speed test on a desktop), truenas cpu usage is up to 30% per esxtop. Running top within truenas at the same time shows no processes consuming any significant cpu.
The only truenas services running are NFS, SMB, Smart and SSH. The only common factor is the network activity and the fact that the nas is on the same subnet, portgroup and vswitch as the client running speed test. Other vm's set up similarly (nextcloud/ubuntu, cyberpower powerpanel/centos, freepbx/ubuntu) don't exhibit this behavior.
I have a few vlans set up for iot purposes. Speedtests performed on different vlans also seem to cause this excessive cpu.
On a whim, I tried to isolate truenas to a separate portgroup bound to the same vswitch. Config looks like this.
Previously, Truenas.nas was part of the LAN_port(2) portgroup. Repeating the speed tests on the network no longer causes high cpu in truenas.
I don't see any negatives from the updated config but am still puzzled¿? Why this is happening in the first place? Network traffic is still being generated on the same vswitch and everything remains fully accessible.
