[pve-devel] lxc container config format

Dietmar Maurer dietmar at proxmox.com
Sat Aug 1 07:05:29 CEST 2015


Hi all,

we currently try to use the original lxc configuration format to
store our container configuration, because
I initially thought we need that to use things like 'lxc-snapshot'.
But it turns out that we make snapshots using our own commands, and we
do not store snapshot config in 'lxc-snapshot' format at all.

For example a simple config looks like:

# cat /etc/pve/lxc/101/config
lxc.include = /usr/share/lxc/config/debian.common.conf
lxc.arch = i386
lxc.cgroup.cpu.cfs_period_us = 100000
lxc.cgroup.cpu.cfs_quota_us = 150000
lxc.cgroup.cpu.shares = 1073
lxc.cgroup.memory.limit_in_bytes = 536870912
lxc.cgroup.memory.memsw.limit_in_bytes = 1073741824
lxc.rootfs = loop:/var/lib/vz/images/101/vm-101-rootfs.raw
lxc.utsname = ct1
pve.disksize = 8
pve.volid = local:101/vm-101-rootfs.raw
lxc.network.type = veth
pve.network.bridge = vmbr0
lxc.network.hwaddr = B6:87:08:C2:93:AB
pve.network.ip = dhcp
lxc.network.name = eth0
lxc.network.veth.pair = veth101.0

As you see, this is clumsy, and can be translated to:

arch: i386
os: debian
cpulimit: 1.5
cpuunits: 1073
rootfs: local:101/vm-101-rootfs.raw,size=8 
hostname: ct1
memory: 512
swap: 512
net0: name=eth0,hwaddr=B6:87:08:C2:93:AB,bridge=vmbr0,ip=dhcp

We already expose this simpler config on the API, for example you can
get it with

# pct config 101

So I wonder if we can/should simply store the configuration using that
simplified format? Any objections?




More information about the pve-devel mailing list