microagent resize

Last updated: 2026-08-15

microagent resize <workspace> --size-mib <n> [options]

resize grows or shrinks a stopped workspace’s rootfs disk in place. The workspace must be halted or stopped, and must have no snapshots - a snapshot’s machine state captures device geometry, and a restore would replace the resized disk with the snapshot’s own anyway. Growing extends the disk immediately; shrinking refuses when the target is smaller than the filesystem’s own reported usage. For a named volume’s disk, use volume resize instead.

Both directions run entirely on the host, offline: the rootfs’s ext4 filesystem is grown or shrunk with resize2fs, and a shrink runs e2fsck -f first (resize2fs’s own precondition). Nothing in the guest needs to be aware a resize happened; the new size takes effect the next time the workspace starts.

In a terminal, resize reports the validation, filesystem check, disk, filesystem, verification, and publication phases on stderr when the operation takes long enough to notice. Publication is reported only after the workspace manifest records the verified size. JSON and MCP responses contain no terminal progress text.

Examples

Grow a workspace’s disk before installing something large:

microagent halt research
microagent resize research --size-mib 16384
microagent start research

Check the result:

microagent --json resize research --size-mib 16384

{
  "workspace": "research",
  "from_size_mib": 8192,
  "to_size_mib": 16384,
  "usage": {
    "sizeMiB": 16384,
    "fsUsedMiB": 912,
    "fsFreeMiB": 15472,
    "hostAllocatedMiB": 950,
    "usedPercent": 6
  }
}

Flags

Flag	Description
--size-mib <n>	Target rootfs size in MiB
--resize2fs <path>	resize2fs binary path
--backend <name>	Backend identity override
--state-dir <dir>	State directory holding the workspace (default ~/.microagent/)

See global flags for --output/--json.

Exit status

resize exits 0 on success. It exits nonzero when the workspace cannot be found, is running, starting, paused, or quarantined, has one or more snapshots, or when a shrink target is smaller than the filesystem’s own reported usage.

Related

• volume - volume resize for a named volume’s disk
• status - inspect/status report the disk’s provisioned, filesystem-used, and host-allocated size
• snapshot - delete snapshots before resizing the rootfs