you can already install it now by running "apt install pve-kernel-4.15". once it is made the default, you don't need to switch manually as it will be automatically be pulled in by the 'proxmox-ve' meta package.
4.15 based test kernel for PVE 5.x available
- Thread starter fabian
- Start date
you can already install it now by running "apt install pve-kernel-4.15". once it is made the default, you don't need to switch manually as it will be automatically be pulled in by the 'proxmox-ve' meta package.
Sometimes if i say "reboot" in an LXC the whole server did not response anymore. On the clusternode on all VM's, containers and storages there is an questionmark. On journallog there is nothing showy. PCT Command did nothing, only an never ending timeout. Only an reboot did help here.
if you can reproduce this, can you obtain more details such as
- pveversion -v
- container and storage configs
- sys-rq traces of all tasks
- output of an LXC debug log file obtained by starting the container in foreground mode with lxc-start
Sorry @fabian , at this weeks no time for it. I rolled back to kernel 4.10 the little homecluster 
What i can remember
- it crashes most on start/shutdown/reboot
- if crash on the clusternode on every icon there is an questionmark
- no errors in journal
- it was on all containers that i used this time for develop (all had an mountpoint to an local zfsdataset)
What can i do to fix it? A reboot did not help really, the machine is hanging with no timeout on stopping VMs/Container. What works is to search with "ps" for the process and kill it.
Hope this help a little bit.
What i can remember
- it crashes most on start/shutdown/reboot
- if crash on the clusternode on every icon there is an questionmark
- no errors in journal
- it was on all containers that i used this time for develop (all had an mountpoint to an local zfsdataset)
What can i do to fix it? A reboot did not help really, the machine is hanging with no timeout on stopping VMs/Container. What works is to search with "ps" for the process and kill it.
Hope this help a little bit.
You still need the java fix to run manually or via cron. The 4.15 kernel on that matter won't change anything.
Last edited:
Working perfectly on our pure Intel S2600WT servers. The in-tree Intel drivers additionally allow us to run without disabling gro (generic receive offloading), which we had to do on every Proxmox kernel before (only appeared to affect older RHEL5 guests).
because at some point, we need to switch (the new kernel has a lot of issues fixed which the 4.13 kernels still have, and having a kernel that never shows any issues on all possible hardware is an illusion). the above known issue will hopefully be fixed with the next round of kernel upgrades. for the time being, simply boot the old working kernel.
If you have two kernels installed, so one 4.13 and one 4.15, edit the file and update Grub.
Code:
cat /etc/default/grub
...
GRUB_DEFAULT="1>1"
#GRUB_DEFAULT=0
...
Code:
update-grub