That is a sore subject with me. The documentation is lacking to say the least.
Here is what was in the prereq's:
- The BYO storage server with Red Hat Enterprise Linux 7.6 and later, same as
the NetBackup Appliance operating system version. - The BYO storage server with docker installed.
- The docker version must be same as the one in the corresponding official
RHEL version release. You need to install it from the corresponding RHEL
yum source (RHEL extra). - The docker application is included in the environment path.
- The BYO storage server with NFS service installed.
- The BYO storage server with NGINX version installed.
- The NGINX version must be same as the one in the corresponding official
RHEL version release. You need to install it from the corresponding RHEL
yum source (epel). - Ensure that the policycoreutils and policycoreutils-python packages
are installed from the same RHEL yum source (RHEL server) and then run
the following commands:- semanage port -a -t http_port_t -p tcp 10087
- setsebool -P httpd_can_network_connect 1
- Ensure that the /mnt folder on the storage server is not mounted by any
mount points directly. Mount points should be mounted to its subfolders. - Enable the logrotate permission in selinux using the following command:
semanage permissive -a logrotate_t
- For BYO, docker container is used to browse VMDK files. Data related to the
container is stored at the following location: /var/lib/ and requires minimum
20 GB free space.
I believe everything is satisfied.
In my environment, my data network and my storage network are two different network spaces that are not routed. As far as I can tell, NetBackup notifies vCenter that the vSphere host should NFS mount the media server. When I go through the VMware | Recovery | Create Instant Access Recovery virtual machine the ESXi host that is selected is the FQDN for the data network, not the storage network. I can't figure out how to tell NBU to use the storage network for the NFS mount.