README.md

   1 # lxcfs
   2
   3 ## Introduction
   4 LXCFS is a small FUSE filesystem written with the intention of making Linux
   5 containers feel more like a virtual machine. It started as a side-project of
   6 `LXC` but is useable by any runtime.
   7
   8 LXCFS will take care that the information provided by crucial files in `procfs`
   9 such as:
  10
  11 ```
  12 /proc/cpuinfo
  13 /proc/diskstats
  14 /proc/meminfo
  15 /proc/stat
  16 /proc/swaps
  17 /proc/uptime
  18 /proc/slabinfo
  19 /sys/devices/system/cpu
  20 /sys/devices/system/cpu/online
  21 ```
  22
  23 are container aware such that the values displayed (e.g. in `/proc/uptime`)
  24 really reflect how long the container is running and not how long the host is
  25 running.
  26
  27 Prior to the implementation of cgroup namespaces by Serge Hallyn `LXCFS` also
  28 provided a container aware `cgroupfs` tree. It took care that the container
  29 only had access to cgroups underneath it's own cgroups and thus provided
  30 additional safety. For systems without support for cgroup namespaces `LXCFS`
  31 will still provide this feature but it is mostly considered deprecated.
  32
  33 ## Upgrading `LXCFS` without restart
  34
  35 `LXCFS` is split into a shared library (a libtool module, to be precise)
  36 `liblxcfs` and a simple binary `lxcfs`. When upgrading to a newer version of
  37 `LXCFS` the `lxcfs` binary will not be restarted. Instead it will detect that
  38 a new version of the shared library is available and will reload it using
  39 `dlclose(3)` and `dlopen(3)`. This design was chosen so that the fuse main loop
  40 that `LXCFS` uses will not need to be restarted. If it were then all containers
  41 using `LXCFS` would need to be restarted since they would otherwise be left
  42 with broken fuse mounts.
  43
  44 To force a reload of the shared library at the next possible instance simply
  45 send `SIGUSR1` to the pid of the running `LXCFS` process. This can be as simple
  46 as doing:
  47
  48     kill -s USR1 $(pidof lxcfs)
  49
  50 ### musl
  51
  52 To achieve smooth upgrades through shared library reloads `LXCFS` also relies
  53 on the fact that when `dlclose(3)` drops the last reference to the shared
  54 library destructors are run and when `dlopen(3)` is called constructors are
  55 run. While this is true for `glibc` it is not true for `musl` (See the section
  56 [Unloading libraries](https://wiki.musl-libc.org/functional-differences-from-glibc.html).).
  57 So users of `LXCFS` on `musl` are advised to restart `LXCFS` completely and all
  58 containers making use of it.
  59
  60 ## Building
  61 Build lxcfs as follows:
  62
  63     yum install fuse fuse-lib fuse-devel
  64     git clone git://github.com/lxc/lxcfs
  65     cd lxcfs
  66     ./bootstrap.sh
  67     ./configure
  68     make
  69     make install
  70
  71 ## Usage
  72 The recommended command to run lxcfs is:
  73
  74     sudo mkdir -p /var/lib/lxcfs
  75     sudo lxcfs /var/lib/lxcfs
  76
  77 A container runtime wishing to use `LXCFS` should then bind mount the
  78 approriate files into the correct places on container startup.
  79
  80 ### LXC
  81 In order to use lxcfs with systemd-based containers, you can either use
  82 LXC 1.1 in which case it should work automatically, or otherwise, copy
  83 the `lxc.mount.hook` and `lxc.reboot.hook` files (once built) from this tree to
  84 `/usr/share/lxcfs`, make sure it is executable, then add the
  85 following lines to your container configuration:
  86 ```
  87 lxc.mount.auto = cgroup:mixed
  88 lxc.autodev = 1
  89 lxc.kmsg = 0
  90 lxc.include = /usr/share/lxc/config/common.conf.d/00-lxcfs.conf
  91 ```
  92
  93 ## Using with Docker
  94
  95 ```
  96 docker run -it -m 256m --memory-swap 256m \
  97       -v /var/lib/lxcfs/proc/cpuinfo:/proc/cpuinfo:rw \
  98       -v /var/lib/lxcfs/proc/diskstats:/proc/diskstats:rw \
  99       -v /var/lib/lxcfs/proc/meminfo:/proc/meminfo:rw \
 100       -v /var/lib/lxcfs/proc/stat:/proc/stat:rw \
 101       -v /var/lib/lxcfs/proc/swaps:/proc/swaps:rw \
 102       -v /var/lib/lxcfs/proc/uptime:/proc/uptime:rw \
 103       -v /var/lib/lxcfs/proc/slabinfo:/proc/slabinfo:rw \
 104       ubuntu:18.04 /bin/bash
 105  ```
 106
 107  In a system with swap enabled, the parameter "-u" can be used to set all values in "meminfo" that refer to the swap to 0.
 108
 109  sudo lxcfs -u /var/lib/lxcfs
 110
 111 ## Swap handling
 112 If you noticed LXCFS not showing any SWAP in your container despite
 113 having SWAP on your system, please read this section carefully and look
 114 for instructions on how to enable SWAP accounting for your distribution.
 115
 116 Swap cgroup handling on Linux is very confusing and there just isn't a
 117 perfect way for LXCFS to handle it.
 118
 119 Terminology used below:
 120  - RAM refers to `memory.usage_in_bytes` and `memory.limit_in_bytes`
 121  - RAM+SWAP refers to `memory.memsw.usage_in_bytes` and `memory.memsw.limit_in_bytes`
 122
 123 The main issues are:
 124  - SWAP accounting is often opt-in and, requiring a special kernel boot
 125    time option (`swapaccount=1`) and/or special kernel build options
 126    (`CONFIG_MEMCG_SWAP`).
 127
 128  - Both a RAM limit and a RAM+SWAP limit can be set. The delta however
 129    isn't the available SWAP space as the kernel is still free to SWAP as
 130    much of the RAM as it feels like. This makes it impossible to render
 131    a SWAP device size as using the delta between RAM and RAM+SWAP for that
 132    wouldn't account for the kernel swapping more pages, leading to swap
 133    usage exceeding swap total.
 134
 135  - It's impossible to disable SWAP in a given container. The closest
 136    that can be done is setting swappiness down to 0 which severly limits
 137    the risk of swapping pages but doesn't eliminate it.
 138
 139 As a result, LXCFS had to make some compromise which go as follow:
 140  - When SWAP accounting isn't enabled, no SWAP space is reported at all.
 141    This is simply because there is no way to know the SWAP consumption.
 142    The container may very much be using some SWAP though, there's just
 143    no way to know how much of it and showing a SWAP device would require
 144    some kind of SWAP usage to be reported. Showing the host value would be
 145    completely wrong, showing a 0 value would be equallty wrong.
 146
 147  - Because SWAP usage for a given container can exceed the delta between
 148    RAM and RAM+SWAP, the SWAP size is always reported to be the smaller of
 149    the RAM+SWAP limit or the host SWAP device itself. This ensures that at no
 150    point SWAP usage will be allowed to exceed the SWAP size.
 151
 152  - If the swappiness is set to 0 and there is no SWAP usage, no SWAP is reported.
 153    However if there is SWAP usage, then a SWAP device of the size of the
 154    usage (100% full) is reported. This provides adequate reporting of
 155    the memory consumption while preventing applications from assuming more
 156    SWAP is available.