MaaS 2.9.2 improperly detects spinning disks as SSD's

[David Andruczyk]

When commissioning a host with multiple drives both SSD and rotary, maas is detecting the spinning drives as SSD's which breaks our automation tools which can compare against tags assigned to drives (ssd vs rotary) among other criteria..

This didn't happen in 2.6.x

[Lee Trager (ltrager)]

Could you provide the output of 50-maas-01-commissioning?

[David Andruczyk (dandruczyk)]

--2021-04-26 21:46:47-- http://10.41.16.254:5248/machine-resources/amd64
Connecting to 10.41.16.254:5248... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6480216 (6.2M) [application/octet-stream]
Saving to: ‘/tmp/user_data.sh.29aYa5/scripts/downloads/50-maas-01-commissioning/machine-resources’

     0K .......... .......... .......... .......... .......... 0% 59.3M 0s
    50K .......... .......... .......... .......... .......... 1% 82.4M 0s
   100K .......... .......... .......... .......... .......... 2% 65.6M 0s
   150K .......... .......... .......... .......... .......... 3% 85.3M 0s
   200K .......... .......... .......... .......... .......... 3% 85.0M 0s
   250K .......... .......... .......... .......... .......... 4% 84.0M 0s
   300K .......... .......... .......... .......... .......... 5% 120M 0s
   350K .......... .......... .......... .......... .......... 6% 81.1M 0s
   400K .......... .......... .......... .......... .......... 7% 92.9M 0s
   450K .......... .......... .......... .......... .......... 7% 118M 0s
   500K .......... .......... .......... .......... .......... 8% 120M 0s
   550K .......... .......... .......... .......... .......... 9% 92.0M 0s
   600K .......... .......... .......... .......... .......... 10% 121M 0s
   650K .......... .......... .......... .......... .......... 11% 87.9M 0s
   700K .......... .......... .......... .......... .......... 11% 120M 0s
   750K .......... .......... .......... .......... .......... 12% 115M 0s
   800K .......... .......... .......... .......... .......... 13% 126M 0s
   850K .......... .......... .......... .......... .......... 14% 96.7M 0s
   900K .......... .......... .......... .......... .......... 15% 116M 0s
   950K .......... .......... .......... .......... .......... 15% 119M 0s
  1000K .......... .......... .......... .......... .......... 16% 94.3M 0s
  1050K .......... .......... .......... .......... .......... 17% 119M 0s
  1100K .......... .......... .......... .......... .......... 18% 119M 0s
  1150K .......... .......... .......... .......... .......... 18% 94.0M 0s
  1200K .......... .......... .......... .......... .......... 19% 117M 0s
  1250K .......... .......... .......... .......... .......... 20% 128M 0s
  1300K .......... .......... .......... .......... .......... 21% 89.0M 0s
  1350K .......... .......... .......... .......... .......... 22% 126M 0s
  1400K .......... .......... .......... .......... .......... 22% 118M 0s
  1450K .......... .......... .......... .......... .......... 23% 120M 0s
  1500K .......... .......... .......... .......... .......... 24% 93.9M 0s
  1550K .......... .......... .......... .......... .......... 25% 119M 0s
  1600K .......... .......... .......... .......... .......... 26% 125M 0s
  1650K .......... .......... .......... .......... .......... 26% 118M 0s
  1700K .......... .......... .......... .......... .......... 27% 94.0M 0s
  1750K .......... .......... .......... .......... .......... 28% 119M 0s
  1800K .......... .......... .......... .......... .......... 29% 126M 0s
  1850K .......... .......... .......... .......... .......... 30% 119...

[Sandor Zeestraten (szeestraten)]

Apologies for butting in, but I might have a similar case in https://bugs.launchpad.net/maas/+bug/1921475

[Lee Trager (ltrager)]

@dandruczyk - According to that LXD is not detecting the RPM speed for any drive on your system. MAAS determines whether a disk is an SSD or a rotary disk based on that value.

Could you check if this is fixed in the latest version of LXD? You can do that in a deployed environment or ephemeral with the following

sudo snap refresh --channel=latest/edge lxd
sudo apt-get install -y jq
curl -G --unix-socket "/var/snap/lxd/common/lxd/unix.socket" "lxd/1.0/resources" 2>/dev/null | jq

If RPM is still 0 in the latest version please report it as a bug to the LXD team at https://github.com/lxc/lxd/issues/new

[David Andruczyk (dandruczyk)]

The latest version of the lxd snap still has the same problem.
~# snap info lxd
name: lxd
summary: System container and virtual machine manager
publisher: Canonical✓
store-url: https://snapcraft.io/lxd
contact: https://github.com/lxc/lxd/issues
license: unset
description: |
  **LXD is a system container and virtual machine manager**

  With LXD you can run hundreds of containers of a variety of Linux
  distributions, apply resource limits, pass in directories, USB devices
  or GPUs and setup any network and storage you want.

  LXD containers are lightweight, secure by default and a great
  alternative to running Linux virtual machines.

  If you want to run other Operating Systems or special Linux workloads,
  you can use LXD virtual machines instead

  **Run any Linux distribution you want**

  Pre-made images are available for Ubuntu, Alpine Linux, ArchLinux,
  CentOS, Debian, Fedora, Gentoo, OpenSUSE and more.

  A full list of available images can be found here:
  https://images.linuxcontainers.org

  Can't find the distribution you want? It's easy to make your own images
  too, either using our `distrobuilder` tool or by assembling your own image
  tarball by hand.

  **Containers and VMs at scale**

  LXD is network aware and all interactions go through a simple REST API,
  making it possible to remotely interact with instances on remote
  systems, copying and moving them as you wish.

  Want to go big? LXD also has built-in clustering support,
  letting you turn dozens of servers into one big LXD server.

  **Configuration options**

  Supported options for the LXD snap (`snap set lxd KEY=VALUE`):
     - ceph.builtin: Use snap-specific Ceph configuration [default=false]
     - ceph.external: Use the system's ceph tools (ignores ceph.builtin)
     [default=false]
     - criu.enable: Enable experimental live-migration support
     [default=false]
     - daemon.debug: Increase logging to debug level [default=false]
     - daemon.group: Set group of users that can interact with LXD
     [default=lxd]
     - daemon.preseed: Pass a YAML configuration to `lxd init` on initial
     start
     - daemon.syslog: Send LXD log events to syslog [default=false]
     - lvm.external: Use the system's LVM tools [default=false]
     - lxcfs.pidfd: Start per-container process tracking [default=false]
     - lxcfs.loadavg: Start tracking per-container load average
     [default=false]
     - lxcfs.cfs: Consider CPU shares for CPU usage [default=false]
     - openvswitch.builtin: Run a snap-specific OVS daemon [default=false]
     - shiftfs.enable: Enable shiftfs support [default=auto]

  Documentation: https://linuxcontainers.org/lxd/docs/master/
commands:
  - lxd.benchmark
  - lxd.buginfo
  - lxd.check-kernel
  - lxd.lxc
  - lxd.lxc-to-lxd
  - lxd
  - lxd.migrate
services:
  lxd.activate: oneshot, enabled, inactive
  lxd.daemon: simple, enabled, active
snap-id: J60k4JY0HppjwOjW8dZdYc8obXKxujRu
tracking: latest/edge
refresh-date: today at 12:42 UTC
channels:
  latest/stable: 4.13 2021-04-28 (20211) 72MB -
  latest/candidate: 4.13 2021-04-27 (20211) 72MB -
  latest/beta: ↑
  latest/edge: git-ee6eda2 2021-0...

[David Andruczyk (dandruczyk)]

LXD bug filed https://github.com/lxc/lxd/issues/8721, please do not close this issue until that issue is fully resolved and maas is back to a properly working state.

[David Andruczyk (dandruczyk)]

They closed by bug as a duplicate, but this other bug points it to be a problem with systemd/udev
https://github.com/lxc/lxd/issues/8718

[Jerzy Husakowski (jhusakowski)]

Does this issue still occur on the latest MAAS/LXD? There were some fixes/workarounds landed in LXD that should have prevented this issue, and we'd like to check if that's indeed the case.

[Bill Wear (billwear)]

pinged OP by email on resolution/status.