1SYSTEMD-REPART(8) systemd-repart SYSTEMD-REPART(8)
2
3
4
6 systemd-repart, systemd-repart.service - Automatically grow and add
7 partitions
8
10 systemd-repart [OPTIONS...] [[BLOCKDEVICE]...]
11
12 systemd-repart.service
13
15 systemd-repart grows and adds partitions to a partition table, based on
16 the configuration files described in repart.d(5).
17
18 If invoked with no arguments, it operates on the block device backing
19 the root file system partition of the OS, thus growing and adding
20 partitions of the booted OS image itself. When called in the initial
21 RAM disk it operates on the block device backing /sysroot/ instead,
22 i.e. on the block device the system will soon transition into. The
23 systemd-repart.service service is generally run at boot in the initial
24 RAM disk, in order to augment the partition table of the OS before its
25 partitions are mounted. systemd-repart (mostly) operates in a purely
26 incremental mode: it only grows existing and adds new partitions; it
27 does not shrink, delete or move existing partitions. The service is
28 intended to be run on every boot, but when it detects that the
29 partition table already matches the installed repart.d/*.conf
30 configuration files, it executes no operation.
31
32 systemd-repart is intended to be used when deploying OS images, to
33 automatically adjust them to the system they are running on, during
34 first boot. This way the deployed image can be minimal in size and may
35 be augmented automatically at boot when needed, taking possession of
36 disk space available but not yet used. Specifically the following use
37 cases are among those covered:
38
39 • The root partition may be grown to cover the whole available disk
40 space.
41
42 • A /home/, swap or /srv/ partition can be added.
43
44 • A second (or third, ...) root partition may be added, to cover A/B
45 style setups where a second version of the root file system is
46 alternatingly used for implementing update schemes. The deployed
47 image would carry only a single partition ("A") but on first boot a
48 second partition ("B") for this purpose is automatically created.
49
50 The algorithm executed by systemd-repart is roughly as follows:
51
52 1. The repart.d/*.conf configuration files are loaded and parsed, and
53 ordered by filename (without the directory prefix).
54
55 2. The partition table already existing on the block device is loaded
56 and parsed.
57
58 3. The existing partitions in the partition table are matched up with
59 the repart.d/*.conf files by GPT partition type UUID. The first
60 existing partition of a specific type is assigned the first
61 configuration file declaring the same type. The second existing
62 partition of a specific type is then assigned the second
63 configuration file declaring the same type, and so on. After this
64 iterative assigning is complete any left-over existing partitions
65 that have no matching configuration file are considered "foreign"
66 and left as they are. And any configuration files for which no
67 partition currently exists are understood as a request to create
68 such a partition.
69
70 4. Taking the size constraints and weights declared in the
71 configuration files into account, all partitions that shall be
72 created are now allocated to the disk, taking up all free space,
73 always respecting the size and padding requests. Similar, existing
74 partitions that are determined to grow are grown. New partitions
75 are always appended to the end of the existing partition table,
76 taking the first partition table slot whose index is greater than
77 the indexes of all existing partitions. Partition table slots are
78 never reordered and thus partition numbers are ensured to remain
79 stable. Note that this allocation happens in RAM only, the
80 partition table on disk is not updated yet.
81
82 5. All existing partitions for which configuration files exist and
83 which currently have no GPT partition label set will be assigned a
84 label, either explicitly configured in the configuration or (if
85 that's missing) derived automatically from the partition type. The
86 same is done for all partitions that are newly created. These
87 assignments are done in RAM only, too, the disk is not updated yet.
88
89 6. Similarly, all existing partitions for which configuration files
90 exist and which currently have an all-zero identifying UUID will be
91 assigned a new UUID. This UUID is cryptographically hashed from a
92 common seed value together with the partition type UUID (and a
93 counter in case multiple partitions of the same type are defined),
94 see below. The same is done for all partitions that are created
95 anew. These assignments are done in RAM only, too, the disk is not
96 updated yet.
97
98 7. Similarly, if the disk's volume UUID is all zeroes it is also
99 initialized, also cryptographically hashed from the same common
100 seed value. Also, in RAM only, too.
101
102 8. The disk space assigned to new partitions (i.e. what was previously
103 considered free space but is no longer) is now erased.
104 Specifically, all file system signatures are removed, and if the
105 device supports it the BLKDISCARD I/O control command is issued to
106 inform the hardware that the space is empty now. In addition any
107 "padding" between partitions and at the end of the device is
108 similarly erased.
109
110 9. The new partition table is finally written to disk. The kernel is
111 asked to reread the partition table.
112
113 As exception to the normally strictly incremental operation, when
114 called in a special "factory reset" mode, systemd-repart may also be
115 used to erase existing partitions to reset an installation back to
116 vendor defaults. This mode of operation is used when either the
117 --factory-reset=yes switch is passed on the tool's command line, or the
118 systemd.factory_reset=yes option specified on the kernel command line,
119 or the FactoryReset EFI variable (vendor UUID
120 8cf2644b-4b0b-428f-9387-6d876050dc67) is set to "yes". It alters the
121 algorithm above slightly: between the 3rd and the 4th step above any
122 partition marked explicitly via the FactoryReset= boolean is deleted,
123 and the algorithm restarted, thus immediately re-creating these
124 partitions anew empty.
125
126 Note that systemd-repart only changes partition tables, it does not
127 create or resize any file systems within these partitions. A separate
128 mechanism should be used for that, for example systemd-growfs(8) and
129 systemd-makefs.
130
131 The UUIDs identifying the new partitions created (or assigned to
132 existing partitions that have no UUID yet), as well as the disk as a
133 whole are hashed cryptographically from a common seed value. This seed
134 value is usually the machine-id(5) of the system, so that the machine
135 ID reproducibly determines the UUIDs assigned to all partitions. If the
136 machine ID cannot be read (or the user passes --seed=random, see below)
137 the seed is generated randomly instead, so that the partition UUIDs are
138 also effectively random. The seed value may also be set explicitly,
139 formatted as UUID via the --seed= option. By hashing these UUIDs from a
140 common seed images prepared with this tool become reproducible and the
141 result of the algorithm above deterministic.
142
143 The positional argument should specify the block device to operate on.
144 Instead of a block device node path a regular file may be specified
145 too, in which case the command operates on it like it would if a
146 loopback block device node was specified with the file attached. If
147 --empty=create is specified the specified path is created as regular
148 file, which is useful for generating disk images from scratch.
149
151 The following options are understood:
152
153 --dry-run=
154 Takes a boolean. If this switch is not specified --dry-run=yes is
155 the implied default. Controls whether systemd-repart executes the
156 requested re-partition operations or whether it should only show
157 what it would do. Unless --dry-run=no is specified systemd-repart
158 will not actually touch the device's partition table.
159
160 --empty=
161 Takes one of "refuse", "allow", "require", "force" or "create".
162 Controls how to operate on block devices that are entirely empty,
163 i.e. carry no partition table/disk label yet. If this switch is not
164 specified the implied default is "refuse".
165
166 If "refuse" systemd-repart requires that the block device it shall
167 operate on already carries a partition table and refuses operation
168 if none is found. If "allow" the command will extend an existing
169 partition table or create a new one if none exists. If "require"
170 the command will create a new partition table if none exists so
171 far, and refuse operation if one already exists. If "force" it will
172 create a fresh partition table unconditionally, erasing the disk
173 fully in effect. If "force" no existing partitions will be taken
174 into account or survive the operation. Hence: use with care, this
175 is a great way to lose all your data. If "create" a new loopback
176 file is create under the path passed via the device node parameter,
177 of the size indicated with --size=, see below.
178
179 --discard=
180 Takes a boolean. If this switch is not specified --discard=yes is
181 the implied default. Controls whether to issue the BLKDISCARD I/O
182 control command on the space taken up by any added partitions or on
183 the space in between them. Usually, it's a good idea to issue this
184 request since it tells the underlying hardware that the covered
185 blocks shall be considered empty, improving performance. If
186 operating on a regular file instead of a block device node, a
187 sparse file is generated.
188
189 --size=
190 Takes a size in bytes, using the usual K, M, G, T suffixes, or the
191 special value "auto". If used the specified device node path must
192 refer to a regular file, which is then grown to the specified size
193 if smaller, before any change is made to the partition table. If
194 specified as "auto" the minimal size for the disk image is
195 automatically determined (i.e. the minimal sizes of all partitions
196 are summed up, taking space for additional metadata into account).
197 This switch is not supported if the specified node is a block
198 device. This switch has no effect if the file is already as large
199 as the specified size or larger. The specified size is implicitly
200 rounded up to multiples of 4096. When used with --empty=create this
201 specifies the initial size of the loopback file to create.
202
203 The --size=auto option takes the sizes of pre-existing partitions
204 into account. However, it does not accommodate for partition tables
205 that are not tightly packed: the configured partitions might still
206 not fit into the backing device if empty space exists between
207 pre-existing partitions (or before the first partition) that cannot
208 be fully filled by partitions to grow or create.
209
210 Also note that the automatic size determination does not take files
211 or directories specified with CopyFiles= into account: operation
212 might fail if the specified files or directories require more disk
213 space then the configured per-partition minimal size limit.
214
215 --factory-reset=
216 Takes boolean. If this switch is not specified --factory=reset=no
217 is the implied default. Controls whether to operate in "factory
218 reset" mode, see above. If set to true this will remove all
219 existing partitions marked with FactoryReset= set to yes early
220 while executing the re-partitioning algorithm. Use with care, this
221 is a great way to lose all your data. Note that partition files
222 need to explicitly turn FactoryReset= on, as the option defaults to
223 off. If no partitions are marked for factory reset this switch has
224 no effect. Note that there are two other methods to request factory
225 reset operation: via the kernel command line and via an EFI
226 variable, see above.
227
228 --can-factory-reset
229 If this switch is specified the disk is not re-partitioned. Instead
230 it is determined if any existing partitions are marked with
231 FactoryReset=. If there are the tool will exit with exit status
232 zero, otherwise non-zero. This switch may be used to quickly
233 determine whether the running system supports a factory reset
234 mechanism built on systemd-repart.
235
236 --root=
237 Takes a path to a directory to use as root file system when
238 searching for repart.d/*.conf files and for the machine ID file to
239 use as seed. By default when invoked on the regular system this
240 defaults to the host's root file system /. If invoked from the
241 initial RAM disk this defaults to /sysroot/, so that the tool
242 operates on the configuration and machine ID stored in the root
243 file system later transitioned into itself.
244
245 --seed=
246 Takes a UUID as argument or the special value random. If a UUID is
247 specified the UUIDs to assign to partitions and the partition table
248 itself are derived via cryptographic hashing from it. If not
249 specified it is attempted to read the machine ID from the host (or
250 more precisely, the root directory configured via --root=) and use
251 it as seed instead, falling back to a randomized seed otherwise.
252 Use --seed=random to force a randomized seed. Explicitly specifying
253 the seed may be used to generated strictly reproducible partition
254 tables.
255
256 --pretty=
257 Takes a boolean argument. If this switch is not specified, it
258 defaults to on when called from an interactive terminal and off
259 otherwise. Controls whether to show a user friendly table and
260 graphic illustrating the changes applied.
261
262 --definitions=
263 Takes a file system path. If specified the *.conf files are read
264 from the specified directory instead of searching in
265 /usr/lib/repart.d/*.conf, /etc/repart.d/*.conf,
266 /run/repart.d/*.conf.
267
268 --key-file=
269 Takes a file system path. Configures the encryption key to use when
270 setting up LUKS2 volumes configured with the Encrypt=key-file
271 setting in partition files. Should refer to a regular file
272 containing the key, or an AF_UNIX stream socket in the file system.
273 In the latter case a connection is made to it and the key read from
274 it. If this switch is not specified the empty key (i.e. zero length
275 key) is used. This behaviour is useful for setting up encrypted
276 partitions during early first boot that receive their user-supplied
277 password only in a later setup step.
278
279 --tpm2-device=, --tpm2-pcrs=
280 Configures the TPM2 device and list of PCRs to use for LUKS2
281 volumes configured with the Encrypt=tpm2 option. These options take
282 the same parameters as the identically named options to systemd-
283 cryptenroll(1) and have the same effect on partitions where TPM2
284 enrollment is requested.
285
286 -h, --help
287 Print a short help text and exit.
288
289 --version
290 Print a short version string and exit.
291
292 --no-pager
293 Do not pipe output into a pager.
294
295 --no-legend
296 Do not print the legend, i.e. column headers and the footer with
297 hints.
298
299 --json=MODE
300 Shows output formatted as JSON. Expects one of "short" (for the
301 shortest possible output without any redundant whitespace or line
302 breaks), "pretty" (for a pretty version of the same, with
303 indentation and line breaks) or "off" (to turn off JSON output, the
304 default).
305
307 On success, 0 is returned, a non-zero failure code otherwise.
308
310 systemd(1), repart.d(5), machine-id(5), systemd-cryptenroll(1)
311
312
313
314systemd 248 SYSTEMD-REPART(8)