Search results

+1

seems like solved. perhaps after one of updates. will watch it some more days...

we're always updating nodes asap. but usually ain't able to reboot it for a long time to boot new kernel.

there 1 more vm on this node. its looks absolutely ok

nfs: sas-ssd export /mnt/sas_ssd/ssd-nfs...

qm config 100 agent: 1 balloon: 0 boot: cdn bootdisk: virtio0 cores: 6 hotplug: disk,network,usb,cpu ide2: none,media=cdrom memory: 34816 name: <vm-name> net0: virtio=62:A9:85:05:0E:C4,bridge=vmbr2 numa: 0 ostype: l26 scsihw: virtio-scsi-pci smbios1: uuid=140b3542-e740-4819-8591-6fc977c00a2d...

any help is welcome

one of the vms restarted several times. all i can find in logs for last 3 times reboot: 1. Nov 22 14:58:15 node1 QEMU[1688058]: kvm: ../softmmu/physmem.c:3193: address_space_unmap: Assertion `mr != NULL' failed...

Howdy. We have PVE cluster: pveversion -V proxmox-ve: 7.1-1 (running kernel: 5.11.22-5-pve) pve-manager: 7.1-6 (running version: 7.1-6/4e61e21c) pve-kernel-5.13: 7.1-4 pve-kernel-helper: 7.1-4 pve-kernel-5.11: 7.0-10 pve-kernel-5.4: 6.4-6 pve-kernel-5.13.19-1-pve: 5.13.19-2...

it seems like issue's gone. may be after qemu updates or somehow else.

so what?

nothing helped - changing machine type to q35, disabling qemu agent, changing disk bus/device type to scsi. any more suggestions?

so we'll try proposed config changes asap

experiment with vm clone seems useless - without networking and application work load cloned vm almost don't using memory

we can do it but we have to disable networking for clone.

but anyway there was no problems before pve update to 6 branch on same vm config

ok. its very special vm & we can switch it off to apply config changes not every day. i'll try all that u advising asap & will give feedback here.

inside vm: ps aux | grep qemu root 800 0.0 0.0 22516 2660 ? Ss 02:50 0:18 /usr/sbin/qemu-ga --daemonize -m virtio-serial -p /dev/virtio-ports/org.qemu.guest_agent.0 all other vms also use virtioblk

so migrate this vm to another node & back 2-3 times per week looks like only solution for us just for now

and it grows until node goes to reboot