I have a VM with a basic config:
2 sockets, 4 cores each - x86-64-v3
q35 - 10.0+pve1
UEFI / TPM / Secure boot
Auto partition on system disk / guest tools installed
OS: Alma 10.0
Once it is installed and fully updated before I begin any configuration I use proxmox to take a snapshot and get in the console several warnings with things like "Marking clocksource 'tsc' as unstable because the skew is too large:" and its usually off by 200ms.
This is being caused and reproducible by snapshots in this configuration, the first time a snapshot is taken it produces the below:
2 sockets, 4 cores each - x86-64-v3
q35 - 10.0+pve1
UEFI / TPM / Secure boot
Auto partition on system disk / guest tools installed
OS: Alma 10.0
Once it is installed and fully updated before I begin any configuration I use proxmox to take a snapshot and get in the console several warnings with things like "Marking clocksource 'tsc' as unstable because the skew is too large:" and its usually off by 200ms.
This is being caused and reproducible by snapshots in this configuration, the first time a snapshot is taken it produces the below:
Code:
kernel: clocksource: 'kvm-clock' (not 'tsc') is current clocksource.
Sep 27 17:15:55 kernel: clocksource: Clocksource 'tsc' skewed 244446520 ns (244 ms) over watchdog 'kvm-clock' interval of 495959300 ns (495 ms)
Sep 27 17:15:55 kernel: clocksource: 'tsc' cs_nsec: 740405820 cs_now: 5f318ef6f5 cs_last: 5e930f32c7 mask: ffffffffffffffff
Sep 27 17:15:55 kernel: clocksource: timekeeping watchdog on CPU0: Marking clocksource 'tsc' as unstable because the skew is too large:
Sep 27 17:14:11 kernel: clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x33c4fd1e1b9, max_idle_ns: 440795253705 ns
Sep 27 17:14:11 kernel: clocksource: tsc-early: mask: 0xffffffffffffffff max_cycles: 0x33c4fd1e1b9, max_idle_ns: 440795253705 ns