scripts: Add autocluster-test-ctdb script samba_version.py by Amitay Isaacs <amitay@ozlabs.org> Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Install extra packages for CTDB testing These could go into a "test" role but finer granularity might make thing harder to understand. CTDB really should complete the migration to ss. Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Add preliminary playbook to unpack binary tarball This is an optimisation that allows relevant nodes to have the binary tarball unpacked and bootstrap script run in parallel. Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Conditionally use tarball install in "nas" role This uses the ctdb_config_dir fact from the "ctdb" role. Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Add tbuild node type This node type downloads a tarball and builds clustered Samba. Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Add tarball_common role Given a source tarball: * build a binary tarball * install binary tarball * setup binary install for AD node * setup binary install for NAS node Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Handle new style /etc/nfs.conf This also depends on CTDB's statd-callout to be fixed to get NFS_HOSTNAME using nfsconf. Signed-off-by: Martin Schwenke <martin@meltin.net>
ansible/node: Modernise NFS service management nfs-server and rpc-statd are available on CentOS 7, and CTDB already knows about these service names, so changing these is safe. In CentOS >=8 the rpc-quota service is split from the quota package (required by nfs-utils) into a new quota-rpc package. Signed-off-by: Martin Schwenke <martin@meltin.net>
autocluster: Only retry "create" and "setup" steps during "build" "create" and "setup" are used during development when the failures are usually programming errors rather than intermittent errors, so don't retry when those steps are run separately. Signed-off-by: Martin Schwenke <martin@meltin.net>