+present will be activated immediately.
+
+Increasing the number of active devices in a RAID5 is much more
+effort. Every block in the array will need to be read and written
+back to a new location. From 2.6.17, the Linux Kernel is able to do
+this safely, including restart and interrupted "reshape".
+
+When relocating the first few stripes on a raid5, it is not possible
+to keep the data on disk completely consistent and crash-proof. To
+provide the required safety, mdadm disables writes to the array while
+this "critical section" is reshaped, and takes a backup of the data
+that is in that section. This backup is normally stored in any spare
+devices that the array has, however it can also be stored in a
+separate file specified with the
+.B \-\-backup\-file
+option. If this option is used, and the system does crash during the
+critical period, the same file must be passed to
+.B \-\-assemble
+to restore the backup and reassemble the array.
+
+.SS BITMAP CHANGES
+
+A write-intent bitmap can be added to, or removed from, an active
+array. Either internal bitmaps, or bitmaps stored in a separate file,
+can be added. Note that if you add a bitmap stored in a file which is
+in a filesystem that is on the raid array being affected, the system
+will deadlock. The bitmap must be on a separate filesystem.
+
+.SH INCREMENTAL MODE
+
+.HP 12
+Usage:
+.B mdadm \-\-incremental
+.RB [ \-\-run ]
+.RB [ \-\-quiet ]
+.I component-device
+.HP 12
+Usage:
+.B mdadm \-\-incremental \-\-rebuild
+.HP 12
+Usage:
+.B mdadm \-\-incremental \-\-run \-\-scan
+
+
+.PP
+This mode is designed to be used in conjunction with a device
+discovery system. As devices are found in a system, they can be
+passed to
+.B "mdadm \-\-incremental"
+to be conditionally added to an appropriate array.
+
+.I mdadm
+performs a number of tests to determine if the device is part of an
+array, and which array it should be part of. If an appropriate array
+is found, or can be created,
+.I mdadm
+adds the device to the array and conditionally starts the array.
+
+Note that
+.I mdadm
+will only add devices to an array which were previously working
+(active or spare) parts of that array. It does not currently support
+automatic inclusion of a new drive as a spare in some array.
+
+.B "mdadm \-\-incremental"
+requires a bug-fix in all kernels through 2.6.19.
+Hopefully, this will be fixed in 2.6.20; alternately, apply the patch
+which is included with the mdadm source distribution. If
+.I mdadm
+detects that this bug is present, it will abort any attempt to use
+.BR \-\-incremental .
+
+The tests that
+.I mdadm
+makes are as follow:
+.IP +
+Is the device permitted by
+.BR mdadm.conf ?
+That is, is it listed in a
+.B DEVICES
+line in that file. If
+.B DEVICES
+is absent then the default it to allow any device. Similar if
+.B DEVICES
+contains the special word
+.B partitions
+then any device is allowed. Otherwise the device name given to
+.I mdadm
+must match one of the names or patterns in a
+.B DEVICES
+line.
+
+.IP +
+Does the device have a valid md superblock. If a specific metadata
+version is request with
+.B \-\-metadata
+or
+.B \-e
+then only that style of metadata is accepted, otherwise
+.I mdadm
+finds any known version of metadata. If no
+.I md
+metadata is found, the device is rejected.
+
+.IP +
+Does the metadata match an expected array?
+The metadata can match in two ways. Either there is an array listed
+in
+.B mdadm.conf
+which identifies the array (either by UUID, by name, by device list,
+or by minor-number), or the array was created with a
+.B homehost
+specified and that
+.B homehost
+matches the one in
+.B mdadm.conf
+or on the command line.
+If
+.I mdadm
+is not able to positively identify the array as belonging to the
+current host, the device will be rejected.
+
+.IP +
+.I mdadm
+keeps a list of arrays that it has partially assembled in
+.B /var/run/mdadm/map
+(or
+.B /var/run/mdadm.map
+if the directory doesn't exist). If no array exists which matches
+the metadata on the new device,
+.I mdadm
+must choose a device name and unit number. It does this based on any
+name given in
+.B mdadm.conf
+or any name information stored in the metadata. If this name
+suggests a unit number, that number will be used, otherwise a free
+unit number will be chosen. Normally
+.I mdadm
+will prefer to create a partitionable array, however if the
+.B CREATE
+line in
+.B mdadm.conf
+suggests that a non-partitionable array is preferred, that will be
+honoured.
+
+.IP +
+Once an appropriate array is found or created and the device is added,
+.I mdadm
+must decide if the array is ready to be started. It will
+normally compare the number of available (non-spare) devices to the
+number of devices that the metadata suggests need to be active. If
+there are at least that many, the array will be started. This means
+that if any devices are missing the array will not be restarted.
+
+As an alternative,
+.B \-\-run
+may be passed to
+.B mdadm
+in which case the array will be run as soon as there are enough
+devices present for the data to be accessible. For a raid1, that
+means one device will start the array. For a clean raid5, the array
+will be started as soon as all but one drive is present.
+
+Note that neither of these approaches is really ideal. If it can
+be known that all device discovery has completed, then
+.br
+.B " mdadm \-IRs"
+.br
+can be run which will try to start all arrays that are being
+incrementally assembled. They are started in "read-auto" mode in
+which they are read-only until the first write request. This means
+that no metadata updates are made and no attempt at resync or recovery
+happens. Further devices that are found before the first write can
+still be added safely.