In these notes the node to be added is cld-np-19.cloud.pd.infn.it

Disable the node so that no new VMs will be scheduled on this host. Instructions: Disable/Enable a compute node howto

Migrate the instances on this compute node to other hypervisors. Instructions: Migrate an instance howto (warning: there was once a problem with live-block migration; we therefore decided to always use cold migration)


Just in case save the file /etc/iscsi/initiatorname.iscsi and the output of "ifconfig" command in cld-ctrl-01:/root/Backup/<hostname>


Reinstall the host with CentOS8 stream using Foreman and move the host to the 'hosts_all' hostgroup

As Operating System choose: CentOS Stream 8

As Media choose: CentOS-Stream-local

As Partition table select

  • "Kickstart default - swap 128 GB" (if this machine has 2 disks: one for the OS and one for the VM ephemeral storage). WARNING this will scratch also the second disk. Chidere a Traldi cosa usare per non scratchare il secondo disco
  • "Kickstart default - swap 128 GB 1 disk 2 parts" (if this machine has only one disk for both the OS and the VM ephemeral storage


Once the machine is reinstalled make sure that SElinux is disabled

[root@cld-np-19 ~]# getenforce 
Disabled


Install OpenManage, if this is a DELL host. Instructions: OpenManage howto



Configure the data network (and, for Unipd nodes, also the storage one), e.g.:



nmcli con mod eno3 ipv4.method manual ipv4.addr "192.168.61.129/24" 
nmcli con mod eno3 802-3-ethernet.mtu 9000 
nmcli con up eno3
nmcli con mod eno3 connection.autoconnect yes



Stop puppet:


systemctl stop puppet

In foreman move the host under the ComputeC8CloudVenetoProdTrain hostgroup


Run puppet manually:

puppet agent -t


If this host has 2 disks (one for the OS and one for the ephemeral storage disk), move /var/lib/nova/instances to the second disk


Start and enablepuppet:

systemctl start puppet; systemctl enable puppet

Verify that a ssh from root@cld-log to root@cld-np-19 works and from  nagios@cld-nagios to nagios@cld-np-19  without requiring password

Reboot the node.

Enable the host. From cld-ctrl-01 after having sourced admin-openrc.sh:

openstack compute service set --enable cld-np-19.cloud.pd.infn.it nova-compute


Force the execution of the "VM network and volume" nagios sensor

  • No labels