RGW daemon not start

[sattha-p]

I try vagrant up from this below config with VirtualBox. However RGW is not start. not sure what did I do wrong?

Vagrant.configure("2") do |config|
  config.vm.hostname = "ceph-17.2"
  config.vm.box = "alvistack/ceph-17.2"

  config.vm.provider :libvirt do |libvirt|
    libvirt.cpu_mode = "host-passthrough"
    libvirt.cpus = 2
    libvirt.disk_bus = "virtio"
    libvirt.disk_driver :cache => "writeback"
    libvirt.driver = "kvm"
    libvirt.memory = 8192
    libvirt.memorybacking :access, :mode => "shared"
    libvirt.nested = true
    libvirt.nic_model_type = "virtio"
    libvirt.storage :file, bus: "virtio", cache: "writeback"
    libvirt.video_type = "virtio"
  end

  config.vm.provider :virtualbox do |virtualbox|
    virtualbox.cpus = 2
    virtualbox.memory = 8192
    virtualbox.customize ["modifyvm", :id, "--cpu-profile", "host"]
    virtualbox.customize ["modifyvm", :id, "--nested-hw-virt", "on"]
  end

end

● [email protected] - Ceph rados gateway
     Loaded: loaded (/lib/systemd/system/[email protected]; enabled; vendor>
     Active: failed (Result: exit-code) since Fri 2022-05-13 02:49:13 UTC; 7min>
    Process: 28310 ExecStart=/usr/bin/radosgw -f --cluster ${CLUSTER} --name cl>
   Main PID: 28310 (code=exited, status=1/FAILURE)

May 13 02:49:13 ceph-17 systemd[1]: [email protected]: Scheduled>
May 13 02:49:13 ceph-17 systemd[1]: Stopped Ceph rados gateway.
May 13 02:49:13 ceph-17 systemd[1]: [email protected]: Start req>
May 13 02:49:13 ceph-17 systemd[1]: [email protected]: Failed wi>
May 13 02:49:13 ceph-17 systemd[1]: Failed to start Ceph rados gateway.

[hswong3i]

Hopefully 41bc385 and its new vagrant box could solve this issue.

1. Vagrant + VirtualBox just recently support additional disk with experimental flag; without such disk Ceph OSD could not setup correctly and so RGW may also failed.

In case of Vagrant + Libvirt such support is much easy and out of the box.

2. Vagrant + VirtualBox hard coded first network interface as NAT (not NAT Network), so we could only add additional private interface for multi node interconnectivity, but then our Ansible variables also need to update with such eth1 for all interface related.

In case of Vagrant + Libvirt such support again default with NAT Network so much easy.

3. I have update the README.md example and also provide a sample Vagrantfile, hopefully helpful ;-)