So I've two screenshots. First one is right before the panic, second is it.
![]()
![]()
Please "report-a-bug".... especially if its replicable.
So I've two screenshots. First one is right before the panic, second is it.
![]()
![]()
So I've two screenshots. First one is right before the panic, second is it.
![]()
![]()
No obvious answers but would try the following:Hi,
just updated from TrueNAS 13.0-U1.1. Sadly U2 doesn't come up completely, I had to boot into U1.1 environment again. I tried twice, but U2 always makes the kernel panic.
Unfortunately I just don't know how to provide details. How can I help here?
Environment: TrueNAS 13.0-U1.1 in KVM on Proxmox. I'm using this environment since 12.0-U7 and updated always went smooth.
Bye,
Chris
Hi,
I have the same issue on Proxmox 6.4-13, running the update from TrueNAS-13.0-U1.1.
Getting the same kernel panic and switching back to U1.1 is working great.
CPU is configured as "2 (1 sockets, 2 cores) [host] [cpuunits=1000]".
Thanks Samiel, I'll try!core_pcpu_init() changed in upstream FreeBSD. Try configuring your CPU as 2 (2 sockets, 1 core).
Do you have any reference for the change of `core_pcpu_init()` in the FreeBSD code? Would be good to know.
OK, the panic happened in core_pcpu_init() when trying to load the kernel CPU initialization routine. Typically a problem here is due to the hypervisor reporting the wrong number of cores to TrueNAS. How many cores have you allocated to TrueNAS in ProxMox? How many physical cores do you have in your Ryzen CPU?
I have a very similar setup but with Intel Xeon CPU E5-2620 v4 CPU, 4 vCPUs, 16GB RAM for the TrueNAS VM and saw exactly the same problem when I tried to update from 13.0-U1.1 to -U2 yesterday.To answer all questions:
1. I didn't upgrade the pool since installing 13.0
2. I did a successful scrub.
3. The system has 24GB of RAM.
4. The system doesn't use a thumb drive but a virtual 10GB harddisk.
The TrueNAS VM gets two cores in one socket from Proxmox. The host CPU is a Xeon E5-2630 v2 (12 core). Virtualized CPU type is equal real CPU.
In the meantime I trimmed the boot environment to one (U1.1) before the update and switched from above CPU configuration to 2 sockets, 1 cores and also 1 sockets, 1 cores but behaviour stayed the same.
I don't use any Jails or VM inside TrueNAS.
So I just created https://ixsystems.atlassian.net/browse/NAS-117975
Regards,
Chris
Changing the CPU model from host to kvm64 will make the system boot.I have a very similar setup but with Intel Xeon CPU E5-2620 v4 CPU, 4 vCPUs, 16GB RAM for the TrueNAS VM and saw exactly the same problem when I tried to update from 13.0-U1.1 to -U2 yesterday.
next nightly build
VNET and Berkely Packet Filter are enabled. VNET interface to auto, manually configured 172.16.2.0/24 address with default route of 172.16.2.1 and no IPv6. Network Properties with vnet0:bridge0 and ip4.saddrsel and ip6.saddrsel selected. Also checked are allow_set_hostname, allow_raw_sockets and allow_mlock.How is your jail networking set up?
1. update your jailsVNET and Berkely Packet Filter are enabled. VNET interface to auto, manually configured 172.16.2.0/24 address with default route of 172.16.2.1 and no IPv6. Network Properties with vnet0:bridge0 and ip4.saddrsel and ip6.saddrsel selected. Also checked are allow_set_hostname, allow_raw_sockets and allow_mlock.
It's been like this forever. Nothing different except for the update to TrueNAS.
All the jails run 12.2-RELEASE-p13 except one on 12.3-RELEASE.