GMIRROR(8) FreeBSD System Manager's Manual GMIRROR(8)
NAME
gmirror - control utility for mirrored devices
SYNOPSIS
To compile GEOM_MIRROR into your kernel, add the following lines to your
kernel configuration file:
options GEOM_MIRROR
Alternatively, to load the GEOM_MIRROR module at boot time, add the
following line to your loader.conf(5):
geom_mirror_load="YES"
Usage of the gmirror utility:
gmirror label [-Fhnv] [-b balance] [-s slice] name prov ...
gmirror clear [-v] prov ...
gmirror create [-Fnv] [-b balance] [-s slice] name prov ...
gmirror configure [-adfFhnv] [-b balance] [-s slice] name
gmirror configure [-v] -p priority name prov
gmirror rebuild [-v] name prov ...
gmirror resize [-v] [-s size] name
gmirror insert [-hiv] [-p priority] name prov ...
gmirror remove [-v] name prov ...
gmirror activate [-v] name prov ...
gmirror deactivate [-v] name prov ...
gmirror destroy [-fv] name ...
gmirror forget [-v] name ...
gmirror stop [-fv] name ...
gmirror dump prov ...
gmirror list
gmirror status
gmirror load
gmirror unload
DESCRIPTION
The gmirror utility is used for mirror (RAID1) configurations. After a
mirror's creation, all components are detected and configured
automatically. All operations like failure detection, stale component
detection, rebuild of stale components, etc. are also done automatically.
The gmirror utility uses on-disk metadata (stored in the provider's last
sector) to store all needed information. Since the last sector is used
for this purpose, it is possible to place a root file system on a mirror.
The first argument to gmirror indicates an action to be performed:
label Create a mirror. The order of components is important,
because a component's priority is based on its position
(starting from 0 to 255). The component with the biggest
priority is used by the prefer balance algorithm and is also
used as a master component when resynchronization is needed,
e.g. after a power failure when the device was open for
writing.
Additional options include:
-b balance Specifies balance algorithm to use, one of:
load Read from the component with the
lowest load. This is the default
balance algorithm.
prefer Read from the component with the
biggest priority.
round-robin Use round-robin algorithm when
choosing component to read.
split Split read requests, which are
bigger than or equal to slice size
on N pieces, where N is the number
of active components.
-F Do not synchronize after a power failure or
system crash. Assumes device is in consistent
state.
-h Hardcode providers' names in metadata.
-n Turn off autosynchronization of stale
components.
-s slice When using the split balance algorithm and an
I/O READ request is bigger than or equal to
this value, the I/O request will be split into
N pieces, where N is the number of active
components. Defaults to 4096 bytes.
clear Clear metadata on the given providers.
create Similar to label, but creates mirror without storing on-disk
metadata in last sector. This special "manual" operation
mode assumes some external control to manage mirror detection
after reboot, device hot-plug and other external events.
configure Configure the given device.
Additional options include:
-a Turn on autosynchronization of stale
components.
-b balance Specifies balance algorithm to use.
-d Do not hardcode providers' names in metadata.
-f Synchronize device after a power failure or
system crash.
-F Do not synchronize after a power failure or
system crash. Assumes device is in consistent
state.
-h Hardcode providers' names in metadata.
-n Turn off autosynchronization of stale
components.
-p priority Specifies priority for the given component
prov.
-s slice Specifies slice size for split balance
algorithm.
rebuild Rebuild the given mirror components forcibly. If
autosynchronization was not turned off for the given device,
this command should be unnecessary.
resize Change the size of the given mirror.
Additional options include:
-s size New size of the mirror is expressed in logical
block numbers. This option can be omitted, then
it will be automatically calculated to maximum
available size.
insert Add the given component(s) to the existing mirror.
Additional options include:
-h Hardcode providers' names in metadata.
-i Mark component(s) as inactive immediately
after insertion.
-p priority Specifies priority of the given component(s).
remove Remove the given component(s) from the mirror and clear
metadata on it.
activate Activate the given component(s), which were marked as
inactive before.
deactivate Mark the given component(s) as inactive, so it will not be
automatically connected to the mirror.
destroy Stop the given mirror and clear metadata on all its
components.
Additional options include:
-f
Stop the given mirror even if it is opened.
forget Forget about components which are not connected. This
command is useful when a disk has failed and cannot be
reconnected, preventing the remove command from being used to
remove it.
stop Stop the given mirror.
Additional options include:
-f
Stop the given mirror even if it is opened.
dump Dump metadata stored on the given providers.
list See geom(8).
status See geom(8).
load See geom(8).
unload See geom(8).
Additional options include:
-v
Be more verbose.
EXIT STATUS
Exit status is 0 on success, and 1 if the command fails.
EXAMPLES
Use 3 disks to setup a mirror. Choose split balance algorithm, split
only requests which are bigger than or equal to 2kB. Create file system,
mount it, then unmount it and stop device:
gmirror label -v -b split -s 2048 data da0 da1 da2
newfs /dev/mirror/data
mount /dev/mirror/data /mnt
...
umount /mnt
gmirror stop data
gmirror unload
Create a mirror on disk with valid data (note that the last sector of the
disk will be overwritten). Add another disk to this mirror, so it will
be synchronized with existing disk:
gmirror label -v -b round-robin data da0
gmirror insert data da1
Create a mirror, but do not use automatic synchronization feature. Add
another disk and rebuild it:
gmirror label -v -n -b load data da0 da1
gmirror insert data da2
gmirror rebuild data da2
One disk failed. Replace it with a brand new one:
gmirror forget data
gmirror insert data da1
Create a mirror, deactivate one component, do the backup and connect it
again. It will not be resynchronized, if there is no need to do so
(there were no writes in the meantime):
gmirror label data da0 da1
gmirror deactivate data da1
dd if=/dev/da1 of=/backup/data.img bs=1m
gmirror activate data da1
SYSCTL VARIABLES
The following sysctl(8) variables can be used to configure behavior for
all mirrors.
kern.geom.mirror.debug
Control the verbosity of kernel logging related to mirrors. A
value larger than 0 will enable debug logging.
kern.geom.mirror.timeout
The amount of time, in seconds, to wait for all copies of a
mirror to appear before starting the mirror. Disks that appear
after the mirror has been started are not automatically added to
the mirror.
kern.geom.mirror.idletime
The amount of time, in seconds, which must elapse after the last
write to a mirror before that mirror is marked clean. Clean
mirrors do not need to be synchronized after a power failure or
system crash. A small value may result in frequent overwrites of
the disks' metadata sectors, and thus may reduce the longevity of
the disks.
kern.geom.mirror.disconnect_on_failure
Determine whether a disk is automatically removed from its mirror
when an I/O request to that disk fails.
kern.geom.mirror.sync_requests
The number of parallel I/O requests used while synchronizing a
mirror. This parameter may only be configured as a
loader.conf(5) tunable.
kern.geom.mirror.sync_update_period
The period, in seconds, at which a synchronizing mirror's
metadata is updated. Periodic updates are used to record a
synchronization's progress so that an interrupted synchronization
may be resumed starting at the recorded offset, rather than at
the beginning. A smaller value results in more accurate progress
tracking, but also increases the number of non-sequential writes
to the disk being synchronized. If the sysctl value is 0, no
updates are performed until the synchronization is complete.
NOTES
Doing kernel dumps to gmirror providers is possible, but some conditions
have to be met. First of all, a kernel dump will go only to one
component and gmirror always chooses the component with the highest
priority. Reading a dump from the mirror on boot will only work if the
prefer balance algorithm is used (that way gmirror will read only from
the component with the highest priority). If you use a different balance
algorithm, you should create an rc(8) script that sets the balance
algorithm to prefer, for example with the following command:
gmirror configure -b prefer data
Make sure that rcorder(8) schedules the new script before savecore(8).
The desired balance algorithm can be restored later on by placing the
following command in rc.local(8):
gmirror configure -b round-robin data
The decision which component to choose for dumping is made when dumpon(8)
is called. If on the next boot a component with a higher priority will
be available, the prefer algorithm will choose to read from it and
savecore(8) will find nothing. If on the next boot a component with the
highest priority will be synchronized, the prefer balance algorithm will
read from the next one, thus will find nothing there.
SEE ALSO
geom(4), dumpon(8), geom(8), mount(8), newfs(8), savecore(8), sysctl(8),
umount(8)
HISTORY
The gmirror utility appeared in FreeBSD 5.3.
AUTHORS
Pawel Jakub Dawidek <pjd@FreeBSD.org>
BUGS
There should be a way to change a component's priority inside a running
mirror.
There should be a section with an implementation description.
FreeBSD 15.1-STABLE-HBSD January 23, 2025 GMIRROR(8)