siderolabs
Enable "lvcreate --type cache"
I believe the following line is responsible for dm-cache not being supported. Please enable it.
https://github.com/siderolabs/pkgs/blob/fadf1e22a263b3429fa8fd540b4ff5a71ce8ded2/lvm2/pkg.yaml#L30
I don't think it would be enough, as it requires extra tools, e.g.:
│ + # cache_check_executable = "/usr/sbin/cache_check"
Please note that as long as Linux kernel supports it, you can bring your tools in a container, you don't have to use host tools.
For some reason Talos (1.10.1) doesn't pick up externally created (booted into a live CD of some other Linux distro, tried several, used arch doc on LVM for reference) LVM RAID1+cache volumes. Once I do lvconvert --uncache vg0/raid0 in that Live CD and reboot into Talos, the volumes are picked up.
With cache:
talosctl get discoveredvolumes
NODE NAMESPACE TYPE ID VERSION TYPE SIZE DISCOVERED LABEL PARTITIONLABEL
192.168.56.101 runtime DiscoveredVolume loop0 1 disk 78 MB squashfs
192.168.56.101 runtime DiscoveredVolume nvme0n1 1 disk 22 GB gpt
192.168.56.101 runtime DiscoveredVolume nvme0n1p1 1 partition 1.2 GB vfat EFI EFI
192.168.56.101 runtime DiscoveredVolume nvme0n1p2 1 partition 1.0 MB talosmeta META
192.168.56.101 runtime DiscoveredVolume nvme0n1p3 1 partition 105 MB xfs STATE STATE
192.168.56.101 runtime DiscoveredVolume nvme0n1p4 1 partition 2.0 GB xfs EPHEMERAL EPHEMERAL
192.168.56.101 runtime DiscoveredVolume nvme0n1p5 1 partition 18 GB lvm2-pv ffOeCQ-6JDI-DOWP-aYHn-eiJ3-4wSe-PN0paA
192.168.56.101 runtime DiscoveredVolume sda 1 disk 107 GB lvm2-pv uLGhx4-meoB-vzx2-rzzx-OXhL-yDtc-z4NjEc
192.168.56.101 runtime DiscoveredVolume sdb 1 disk 107 GB lvm2-pv NZ550X-dfv3-o2oh-uhE4-cJoe-d9AT-iF6gOC
without cache:
talosctl get discoveredvolumes
NODE NAMESPACE TYPE ID VERSION TYPE SIZE DISCOVERED LABEL PARTITIONLABEL
192.168.56.101 runtime DiscoveredVolume dm-0 1 disk 4.2 MB
192.168.56.101 runtime DiscoveredVolume dm-1 1 disk 107 GB xfs
192.168.56.101 runtime DiscoveredVolume dm-2 1 disk 4.2 MB
192.168.56.101 runtime DiscoveredVolume dm-3 1 disk 107 GB xfs
192.168.56.101 runtime DiscoveredVolume dm-4 1 disk 107 GB xfs
192.168.56.101 runtime DiscoveredVolume loop0 1 disk 78 MB squashfs
192.168.56.101 runtime DiscoveredVolume nvme0n1 1 disk 22 GB gpt
192.168.56.101 runtime DiscoveredVolume nvme0n1p1 1 partition 1.2 GB vfat EFI EFI
192.168.56.101 runtime DiscoveredVolume nvme0n1p2 1 partition 1.0 MB talosmeta META
192.168.56.101 runtime DiscoveredVolume nvme0n1p3 1 partition 105 MB xfs STATE STATE
192.168.56.101 runtime DiscoveredVolume nvme0n1p4 1 partition 2.0 GB xfs EPHEMERAL EPHEMERAL
192.168.56.101 runtime DiscoveredVolume nvme0n1p5 1 partition 18 GB lvm2-pv ffOeCQ-6JDI-DOWP-aYHn-eiJ3-4wSe-PN0paA
192.168.56.101 runtime DiscoveredVolume sda 1 disk 107 GB lvm2-pv uLGhx4-meoB-vzx2-rzzx-OXhL-yDtc-z4NjEc
192.168.56.101 runtime DiscoveredVolume sdb 1 disk 107 GB lvm2-pv NZ550X-dfv3-o2oh-uhE4-cJoe-d9AT-iF6gOC
(Notice the dm-* not present with cache)
Trying to manage volumes from a Docker (e.g. debian) in Talos fail due to inability to access udev socket.
I believe I have seen some error messages in dmesg I could attribute to inability to deal with cache, but now I cannot reproduce that.
P.S. All the relevant kernel modules (dm-* just in case) are loaded:
# Configures the kernel.
kernel:
# Kernel modules to load.
modules:
- name: dm-cache
- name: dm-cache-smq
- name: dm-raid
- name: dm-thin-pool