These limitations in Bare Metal Orchestrator are not known issues, but can cause unexpected problems.
PowerEdge RAID Controller (PERC) H755 does not provide VMware ESXi with the correct target numbers for virtual disks
If a user provides a volume name under RAID in the installVolumeTypeOrder
field, operating system installation fails because the PERC H755 does not provide VMware ESXi with the correct target numbers for virtual disks.
No DHCP support for IPv6
DHCP support is only available for IPv4 addresses.
Operating systems using static IP addresses
Hardware profiles can deploy operating systems to servers using static IP addresses, which can result in operating systems using duplicate IP addresses. The Bare Metal Orchestrator web interface warns you if you attempt to do with this with VMware ESXi, but there is no warning from the command line interface.
Retrieve operating system inventory for servers with ESXi
Bare Metal Orchestrator supports retrieving operating system inventory only from onboarded servers that have the ESXi operating system installed.
Server operating system ISO filename character restrictions
Before deploying an operating system on servers using Bare Metal Orchestrator, ensure that the ISO filename has at least four hyphens and that the ISO filename does not contain restricted characters ( / < > | : &) for Linux-based filenames. For example, change rhel-8.4-x86_64-dvd.iso to rhel-8.4-x86-64-dvd.iso.
Concurrent Red Hat Enterprise Linux installations time out
The Red Hat Enterprise Linux operating system can only be installed on one server at a time for a site. Otherwise, an operation timeout error occurs.
iDRAC firmware 6.10.30.00 limitation
Servers that have iDRACs with firmware version 6.10.30.00 installed do not work with Bare Metal Orchestrator if they are located behind a firewall or a NAT device. These servers must be on the same local network as the Bare Metal Orchestrator cluster.
Two servers with the same IP address
Two scenarios can occur if two servers have the same IP address.
- If a server goes offline, loses its IP address lease, and comes back online after its IP address is reassigned, the first server will be placed in a failed state.
- If you attempt to automatically onboard a server that uses the same IP address as a currently onboarded server, Bare Metal Orchestrator blocks the attempt and you must manually onboard the new server.
PV backup does not capture audit logs
A PV back up does not back up images and audit logs. Also, scheduled PV backups are not supported.
Applying hardware profiles in quick succession
After applying a hardware profile to a server, wait for the server modifications to be completed before you modify the hardware profile and apply it to another server. Otherwise, you get a 409 conflict error with the message "hardware profile already in progress." The amount of time it takes for modifications to complete depends on the number of operations specified in the hardware profile.
Scheduled backups not supported using the CLI or Web UI
Do not use the scheduled backups feature for Bare Metal Orchestrator release 2.1.2. Currently, the scheduled backup feature captures only ETCD data. Perform a manual backup using the Bare Metal Orchestrator CLI instead. The procedure to manually back up Bare Metal Orchestrator includes a script to capture both ETCD and PostgreSQL data.
Limitation on the number of brownfield servers that can be migrated
You can migrate a maximum of 10 servers per site at a time when performing a data and server migration. If you have more than 100 servers, you should migrate the first 10 servers per site, then repeat the migration process as many times as required for the remaining servers.
High availability cluster limitation
When a control plane node failure occurs, there will be a downtime of five minutes before the GC site gets into the ready state again.