[Bug 2152460] Re: Ceph 20.2.0 - cephadm cluster bootstrap fails due to hardcoded uid and gid

[Luciano Lo Giudice]

This bug is being handled with the SRU for 20.2.1:
https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/2151199

[ Impact ]

 * Some cephadm commands fail due to the tool assuming that the 'ceph' user ID
   and group ID have pre-determined values.
   
 * THe new packages will solve this by fetching the actual values for both
   UID and GID instead of assuming their values.

[ Test Plan ]

 * In order to test this, the functional tests for Canonical's Ceph containers
   repository will be run (https://github.com/canonical/ceph-containers).

 * These tests involve using the cephadm tool for cluster deployment using
   Canonical's custom Ceph images.

 * If the new packages fix this issue, the tests will run to completion.

[ Where problems could occur ]

 * cephadm is used as a simple tool to deploy Ceph clusters. Notably, neither
   the Ceph charms nor Microceph make use of it, but users of Ubuntu can make
   use of it independently.

 * Any problems with cephadm will thus prevent users from deploying their own
   Ceph clusters with the exceptions noted above.

-- 
You received this bug notification because you are a member of Ubuntu
OpenStack, which is subscribed to ceph in Ubuntu.
https://bugs.launchpad.net/bugs/2152460

Title:
  Ceph 20.2.0 - cephadm cluster bootstrap fails due to hardcoded uid and
  gid

Status in ceph package in Ubuntu:
  New

Bug description:
  After installing cephadm on Resolute:

  ii cephadm 20.2.0-0ubuntu2 all

  Trying to bootstrap a cluster:

  $ sudo cephadm bootstrap --mon-ip 10.3.1.190
  Creating directory /etc/ceph for ceph.conf
  Verifying podman|docker is present...
  Verifying lvm2 is present...
  Verifying time synchronization is in place...
  Unit chrony.service is enabled and running
  Repeating the final host check...
  docker (/usr/bin/docker) is present
  systemctl is present
  lvcreate is present
  Unit chrony.service is enabled and running
  Host looks OK
  Cluster fsid: e4638df2-4eaf-11f1-8000-fa163ebbf652
  Verifying IP 10.3.1.190 port 3300 ...
  Verifying IP 10.3.1.190 port 6789 ...
  Mon IP `10.3.1.190` is in CIDR network `10.3.0.0/21`
  Mon IP `10.3.1.190` is in CIDR network `10.3.0.0/21`
  Mon IP `10.3.1.190` is in CIDR network `10.3.0.1/32`
  Mon IP `10.3.1.190` is in CIDR network `10.3.0.1/32`
  Mon IP `10.3.1.190` is in CIDR network `10.3.1.100/32`
  Mon IP `10.3.1.190` is in CIDR network `10.3.1.100/32`
  Internal network (--cluster-network) has not been provided, OSD replication will default to the public_network
  Pulling container image quay.io/ceph/ceph:v20...
  Ceph version: ceph version 20.2.1 (6a49aff47758778a5f5951e731d437c317f72fb2) tentacle (stable)
  Extracting ceph user uid/gid from container image...
  Creating initial keys...
  Creating initial monmap...
  Creating mon...
  Non-zero exit code 1 from install -d -m0770 -o 167 -g 167 /var/run/ceph/e4638df2-4eaf-11f1-8000-fa163ebbf652
  install: stderr install: invalid user: '167'
  RuntimeError: Failed command: install -d -m0770 -o 167 -g 167 /var/run/ceph/e4638df2-4eaf-11f1-8000-fa163ebbf652: install: invalid user: '167'

          ***************
          Cephadm hit an issue during cluster installation. Current cluster files will be deleted automatically.
          To disable this behaviour you can pass the --no-cleanup-on-failure flag. In case of any previous
          broken installation, users must use the following command to completely delete the broken cluster:

          > cephadm rm-cluster --force --zap-osds --fsid <fsid>

          for more information please refer to https://docs.ceph.com/en/latest/cephadm/operations/#purging-a-cluster
          ***************

  Deleting cluster with fsid: e4638df2-4eaf-11f1-8000-fa163ebbf652
  Traceback (most recent call last):
    File "/usr/sbin/cephadm", line 5288, in <module>
      main()
      ~~~~^^
    File "/usr/sbin/cephadm", line 5276, in main
      r = ctx.func(ctx)
    File "/usr/sbin/cephadm", line 2524, in _rollback
      return func(ctx)
    File "/usr/sbin/cephadm", line 434, in _default_image
      return func(ctx)
    File "/usr/sbin/cephadm", line 2684, in command_bootstrap
      make_var_run(ctx, fsid, uid, gid)
      ~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^
    File "/usr/sbin/cephadm", line 508, in make_var_run
      call_throws(ctx, ['install', '-d', '-m0770', '-o', str(uid), '-g', str(gid),
      ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                        '/var/run/ceph/%s' % fsid])
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/lib/python3/dist-packages/cephadmlib/call_wrappers.py", line 307, in call_throws
      raise RuntimeError(
          f'Failed command: {" ".join(command)}: {s}'
      )

  RuntimeError: Failed command: install -d -m0770 -o 167 -g 167
  /var/run/ceph/e4638df2-4eaf-11f1-8000-fa163ebbf652: install: invalid
  user: '167'

  On RHEL, uid/gid is reserved for the ceph user, however we do not have
  that on Ubuntu and this needs to get fixed on the next SRU.

  Thanks,
  Alan

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/2152460/+subscriptions