Disk Manager

#What it does

The Disk Manager lives in the admin panel (port 2087, /admin) under Server Admin → Disk Manager and is available to the server administrator only; resellers, team members and account users cannot see it. The page reads the server's disks, partitions, LVM groups and filesystems and lists what can be done right now under Suggested operations. Tabs: Disks, LVM groups, Home volumes and History. The disk list shows each disk's role (System disk, In use, Unused, Not supported) and, when smartctl is installed, its health.

The feature is part of both Community and Pro. Only moving existing accounts to another disk requires Pro (see below).

#Three situations

#Adding a disk to an LVM volume

When the filesystem to grow (/, /home, or a home volume such as /home2) sits on LVM, the blank disk joins that volume's group. You choose All available space or A specific amount in GB. If the group already has free space, no disk is needed and the volume simply grows into it. Taking the disk out of the group later requires moving data off it (pvmove), which the panel does not do.

#Preparing an extra home volume

A root filesystem that is not on LVM cannot be merged with a new disk, so the blank disk becomes a separate home volume instead. The suggested mount point is the next free /homeN; system directories, paths inside an existing home volume and non-empty directories are rejected. Options:

• Use LVM (recommended): creates a volume group such as bp_home2 with a logical volume named data, so disks added later can be merged into it. With LVM turned off, that is not possible.
• Filesystem: XFS is recommended on the AlmaLinux/Rocky family and ext4 on Debian/Ubuntu; either can be chosen.
• Make it the default home volume for new accounts: when ticked, new accounts are created on this volume.

The disk goes into /etc/fstab by UUID with quota options and nofail: if it fails, the server still boots and only the accounts on that volume become unavailable. While the disk is not mounted, the mount point is made unwritable, so account data never lands on the root disk by mistake. When SELinux is managed, the new volume is labelled like /home.

#Growing an enlarged virtual disk

After you enlarge a disk in your cloud or hypervisor panel, the partition is extended with growpart, then, depending on its contents, pvresize and lvextend -r (LVM), resize2fs (ext4) or xfs_growfs (XFS) run. If the new space does not show up, press Rescan disks first; it makes the kernel re-read disk sizes and pick up new disks. Disks without a rescan interface, such as virtio disks, may still need a restart of the virtual server if the change stays invisible. Since the partition table changes on the running system, taking a snapshot first is recommended.

#Safety model

• Completely blank disks only: no partitions, partition table, filesystem or RAID/LVM signature, not mounted, not swap, not held by another device, and at least 1 GiB. USB, removable, read-only and multipath disks are never used. There is no wipe or reformat option.
• Preview: every operation starts with Preview, which lists the exact commands, the warnings and the estimated sizes; the commands can be copied.
• Confirmation: to start, you type the disk or volume name (e.g. sdb) and enter your panel password (plus the code when two-factor authentication is on). If the disk state changed after the preview, the operation is refused and you are asked to refresh the preview.
• No data-loss operations: shrinking, deleting partitions, converting an existing root to LVM, RAID0 and pooled filesystems are not offered. /etc/fstab only ever gets a line appended; a /etc/fstab.bipanel-<timestamp> backup is taken first and the new file is verified.
• RAID warning: when a volume will span two disks you must tick the warning box. Without RAID, if either disk fails, all data on that volume is lost, so keep regular backups.
• One disk operation runs at a time, and none starts while a home directory move is running. An operation stops at the first error without an automatic rollback, and the log shows which steps completed. Each operation is written to the audit log and the History tab.

#Home volumes and placement

The Home volumes tab lists /home and the added volumes with usage, account count and disk quota state. A filesystem you mounted outside the panel can be added with Register a mounted volume; it must be a separately mounted, writable directory owned by root and not writable by the group or other users. A volume that still holds accounts cannot be unregistered.

Placement of new accounts follows one of two policies: Always use the default volume or Use the volume with the most free space. With more than one usable volume, the server administrator can pick the Home volume when creating an account, or on the command line:

$ bipanel account create ali example.com --home /home2

Accounts created by resellers follow the placement policy. Community allows up to 3 hosting accounts; see Pricing.

#Disk quotas per volume

Quotas are tracked per home volume, and the tab shows a state such as Enforced or Off for each one. On a volume where quotas are not enforced, account disk limits do not apply, and the panel shows the fix commands to run as root on the server. Without the quota tools, quotas are not turned on for an ext4 volume.

#Moving existing accounts (Pro)

Adding a volume does not move existing accounts. In the Pro edition, an account's home directory can be moved to another home volume with Account Move & Rename → Disk move. In Community, existing accounts stay where they are. See Accounts and packages for account operations.

#Disk tools component

Operations rely on lvm2, parted, sgdisk, growpart, xfsprogs, e2fsprogs and smartmontools. The Disk tools component is installed by default on new installations. If something is missing, the page shows a warning and Install disk tools installs the packages in the background with the package manager; sites are not affected.

#Container installations

The Disk Manager works only on a server installation where the panel runs as root. In a Docker or Railway container, disks and volumes are managed by the platform and the page shows as unavailable. See Docker and Railway.