Hi Proxmox VE community, I just updated Proxmox 6 to 7 on a MicroServer Gen8 and had few issues, I'm writing a post to summarize this journey! The issue seems to be that the hardware is considered "old" and software support is starting to fade. Feel free to suggest and partecipate to the discussion.
Slow console
You can detect if this issue is impacting you if the console is so slow that you can't input keystrokes using the iLO KVM or an attached keyboard and monitor.
This issue is known and documented and has to do with hardware-software interrupt coordination (some sort of "bug"):
hpilo kernel messages and Agentless Management Service (amsd) not working
You can detect if this issue is impacting you in two ways:
1) In dmesg/console you have messages like:
2) In iLO under "Subsystems and Devices" (it's in System Information/Summary) Agentless Management Service is marked as "Not available", regardless if the "amsd" daemon is installed, running, etc.
This happens because the latest "amsd" package supporting iLO 4 is the stretch one. Anyway the binary seems to be pretty self-contained and you can safely set the stretch repository in "/etc/apt/sources.list.d/hpe.list":
Issues to be solved
Slow console
You can detect if this issue is impacting you if the console is so slow that you can't input keystrokes using the iLO KVM or an attached keyboard and monitor.
This issue is known and documented and has to do with hardware-software interrupt coordination (some sort of "bug"):
- https://forum.proxmox.com/threads/updated-to-ve-7-0-no-web-gui-dmar-errors-on-console.92022/
- https://bugzilla.kernel.org/show_bug.cgi?id=214795
- https://www.claudiokuenzler.com/blo...ye-boot-freeze-kernel-panic-hp-proliant-dl380
Code:
GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=igfx_off intremap=off"
hpilo kernel messages and Agentless Management Service (amsd) not working
You can detect if this issue is impacting you in two ways:
1) In dmesg/console you have messages like:
Code:
hpilo 0000:01:00.2: Open could not dequeue a packet
hpilo 0000:01:00.2: Closing, but controller still active
2) In iLO under "Subsystems and Devices" (it's in System Information/Summary) Agentless Management Service is marked as "Not available", regardless if the "amsd" daemon is installed, running, etc.
This happens because the latest "amsd" package supporting iLO 4 is the stretch one. Anyway the binary seems to be pretty self-contained and you can safely set the stretch repository in "/etc/apt/sources.list.d/hpe.list":
Code:
#deb https://downloads.linux.hpe.com/SDR/repo/mcp bullseye/current non-free
# Do not go beyond stretch as iLo 4 is no longer supported, eg:
# hpilo 0000:01:00.2: Open could not dequeue a packet
# hpilo 0000:01:00.2: Closing, but controller still active
deb https://downloads.linux.hpe.com/SDR/repo/mcp stretch/current non-free
Issues to be solved
- Long (see below!) delay after LUKS password input and LVM volume mount
Code:
[ 4.649449] random: lvm: uninitialized urandom read (4 bytes read)
[ 177.116567] usb 2-1: new full-speed USB device number 2 using uhci_hcd
- Add your issue (an solution!)