TrueNAS - PROXMOX - iSCSI : Read connection lost

Hi,

I have setup my first TrueNAS.
For now it’s just a POC for me.

It’s has been installed on a desktop machine :

  • Core I7(Gen12)
  • 32 Gig of RAM
  • 500Gb HD for TrueNAS Core
  • 2x2TB SSD for storage
  • TrueNAS Core 13.0-U6.1
  • 1Gb/s NIC

TrueNAS conf :

  • One static IP connected to a switch on port in access mode.(no vlan, trunk, etherchannel etc.)
  • One pool named data (mirrored using the 2 SSD)
  • One vol created on the “data” named pool, named : truenas
  • iSCSI portal configured with target truenas for all initiator, no authentication

Proxmox conf :

  • 2 servers with proxmox 8.1.3
  • Same subnet as TrueNAS
  • iSCSI connection on each to TrueNAS portal
  • One LVM setup on connected LUN

One Proxmox side everything seems to be OK , I have no error. I have been able to create VM stored on truenas vol.

I made disk perf perf test on the VM, and it also seems to be ok or at least very honest.

But on TrueNAS console I have this log that appears endlessly :

[month] [day] [hour] truenas 1 [date] truenas.local ctld [id] - - [some proxmox ip server]: read: connection lost

ex :

JUN 25 05:36:16 truenas 1 2024-06-25T05:36:16.700194-07:00 truenas.local ctld 6401 - - 10.0.0.1: read: connection lost

As vms work fine on tis storage, I don’t understand this message. I someone can help me to understand and get rid of this : thanks in advance.

Proc.

I’m curious about your problem because I have three TrueNAS Scale NASses connected to Proxmox via iSCSI and I have a very particular performance issue. Here are some questions to isolate:

  • Are you using the TheGrandWazoo repo to connect?
  • Did you set the default block size to 16k?

image

You mention:

I made disk perf perf test on the VM

How?

Finally, you’ve perfectly described your environment except your switch architecture. Are you using 10 Gbit/s switches or something else?

My performance problem really only happens when I run PBS, then I get spammed by this:

Jul 12 04:06:12 example.hv.com QEMU[3644222]: kvm: iSCSI GET_LBA_STATUS failed at lba 139804672: SENSE KEY:ILLEGAL_REQUEST(5) ASCQ:INVALID_FIELD_IN_CDB(0x2400)

For the rest of the time the system runs exceptional. But unfortunately after a number of those spamming events Proxmox Backup Server stop working for some VMs.