1TMPFILES.D(5) tmpfiles.d TMPFILES.D(5)
2
6 tmpfiles.d - Configuration for creation, deletion and cleaning of
7 volatile and temporary files
8
10 /etc/tmpfiles.d/*.conf
11 /run/tmpfiles.d/*.conf
12 /usr/lib/tmpfiles.d/*.conf
13 ~/.config/user-tmpfiles.d/*.conf
16 $XDG_RUNTIME_DIR/user-tmpfiles.d/*.conf
17 ~/.local/share/user-tmpfiles.d/*.conf
18 ...
19 /usr/share/user-tmpfiles.d/*.conf
20 #Type Path Mode User Group Age Argument
24 f /file/to/create mode user group - content
25 f+ /file/to/create-or-truncate mode user group - content
26 w /file/to/write-to - - - - content
27 w+ /file/to/append-to - - - - content
28 d /directory/to/create-and-clean-up mode user group cleanup-age -
29 D /directory/to/create-and-remove mode user group cleanup-age -
30 e /directory/to/clean-up mode user group cleanup-age -
31 v /subvolume-or-directory/to/create mode user group cleanup-age -
32 q /subvolume-or-directory/to/create mode user group cleanup-age -
33 Q /subvolume-or-directory/to/create mode user group cleanup-age -
34 p /fifo/to/create mode user group - -
35 p+ /fifo/to/[re]create mode user group - -
36 L /symlink/to/create - - - - symlink/target/path
37 L+ /symlink/to/[re]create - - - - symlink/target/path
38 c /dev/char-device-to-create mode user group - major:minor
39 c+ /dev/char-device-to-[re]create mode user group - major:minor
40 b /dev/block-device-to-create mode user group - major:minor
41 b+ /dev/block-device-to-[re]create mode user group - major:minor
42 C /target/to/create - - - cleanup-age /source/to/copy
43 x /path-or-glob/to/ignore/recursively - - - cleanup-age -
44 X /path-or-glob/to/ignore - - - cleanup-age -
45 r /path-or-glob/to/remove - - - - -
46 R /path-or-glob/to/remove/recursively - - - - -
47 z /path-or-glob/to/adjust/mode mode user group - -
48 Z /path-or-glob/to/adjust/mode/recursively mode user group - -
49 t /path-or-glob/to/set/xattrs - - - - xattrs
50 T /path-or-glob/to/set/xattrs/recursively - - - - xattrs
51 h /path-or-glob/to/set/attrs - - - - file attrs
52 H /path-or-glob/to/set/attrs/recursively - - - - file attrs
53 a /path-or-glob/to/set/acls - - - - POSIX ACLs
54 a+ /path-or-glob/to/append/acls - - - - POSIX ACLs
55 A /path-or-glob/to/set/acls/recursively - - - - POSIX ACLs
56 A+ /path-or-glob/to/append/acls/recursively - - - - POSIX ACLs
57
60 tmpfiles.d configuration files provide a generic mechanism to define
61 the creation of regular files, directories, pipes, and device nodes,
62 adjustments to their access mode, ownership, attributes, quota
63 assignments, and contents, and finally their time-based removal. It is
64 mostly commonly used for volatile and temporary files and directories
65 (such as those located under /run/, /tmp/, /var/tmp/, the API file
66 systems such as /sys/ or /proc/, as well as some other directories
67 below /var/).
68 systemd-tmpfiles(8) uses this configuration to create volatile files
70 and directories during boot and to do periodic cleanup afterwards. See
71 systemd-tmpfiles(8) for the description of
72 systemd-tmpfiles-setup.service, systemd-tmpfiles-clean.service, and
73 associated units.
74 System daemons frequently require private runtime directories below
76 /run/ to store communication sockets and similar. For these, it is
77 better to use RuntimeDirectory= in their unit files (see
78 systemd.exec(5) for details), if the flexibility provided by tmpfiles.d
79 is not required. The advantages are that the configuration required by
80 the unit is centralized in one place, and that the lifetime of the
81 directory is tied to the lifetime of the service itself. Similarly,
82 StateDirectory=, CacheDirectory=, LogsDirectory=, and
83 ConfigurationDirectory= should be used to create directories under
84 /var/lib/, /var/cache/, /var/log/, and /etc/. tmpfiles.d should be
85 used for files whose lifetime is independent of any service or requires
86 more complicated configuration.
87
89 Each configuration file shall be named in the style of package.conf or
90 package-part.conf. The second variant should be used when it is
91 desirable to make it easy to override just this part of configuration.
92 Files in /etc/tmpfiles.d override files with the same name in
94 /usr/lib/tmpfiles.d and /run/tmpfiles.d. Files in /run/tmpfiles.d
95 override files with the same name in /usr/lib/tmpfiles.d. Packages
96 should install their configuration files in /usr/lib/tmpfiles.d. Files
97 in /etc/tmpfiles.d are reserved for the local administrator, who may
98 use this logic to override the configuration files installed by vendor
99 packages. All configuration files are sorted by their filename in
100 lexicographic order, regardless of which of the directories they reside
101 in. If multiple files specify the same path, the entry in the file with
102 the lexicographically earliest name will be applied (note that lines
103 suppressed due to the "!" are filtered before application, meaning
104 that if an early line carries the exclamation mark and is suppressed
105 because of that, a later line matching in path will be applied). All
106 other conflicting entries will be logged as errors. When two lines are
107 prefix path and suffix path of each other, then the prefix line is
108 always created first, the suffix later (and if removal applies to the
109 line, the order is reversed: the suffix is removed first, the prefix
110 later). Lines that take globs are applied after those accepting no
111 globs. If multiple operations shall be applied on the same file (such
112 as ACL, xattr, file attribute adjustments), these are always done in
113 the same fixed order. Except for those cases, the files/directories are
114 processed in the order they are listed.
115 If the administrator wants to disable a configuration file supplied by
117 the vendor, the recommended way is to place a symlink to /dev/null in
118 /etc/tmpfiles.d/ bearing the same filename.
119
121 The configuration format is one line per path, containing type, path,
122 mode, ownership, age, and argument fields. The lines are separated by
123 newlines, the fields by whitespace:
124 #Type Path Mode User Group Age Argument...
126 d /run/user 0755 root root 10d -
127 L /tmp/foobar - - - - /dev/null
128 Fields may contain C-style escapes. With the exception of the seventh
130 field (the "argument") all fields may be enclosed in quotes. Note that
131 any whitespace found in the line after the beginning of the argument
132 field will be considered part of the argument field. To begin the
133 argument field with a whitespace character, use C-style escapes (e.g.
134 "\x20").
135 Type
137 The type consists of a single letter and optionally one or more
138 modifier characters: a plus sign ("+"), exclamation mark ("!"), minus
139 sign ("-"), equals sign ("="), tilde character ("~") and/or caret
140 ("^").
141 The following line types are understood:
143 f, f+
145 f will create a file if it does not exist yet. If the argument
146 parameter is given and the file did not exist yet, it will be
147 written to the file. f+ will create or truncate the file. If the
148 argument parameter is given, it will be written to the file. Does
149 not follow symlinks.
150 w, w+
152 Write the argument parameter to a file, if the file exists. If
153 suffixed with +, the line will be appended to the file. If your
154 configuration writes multiple lines to the same file, use w+. Lines
155 of this type accept shell-style globs in place of normal path
156 names. The argument parameter will be written without a trailing
157 newline. C-style backslash escapes are interpreted. Follows
158 symlinks.
159 d
161 Create a directory. The mode and ownership will be adjusted if
162 specified. Contents of this directory are subject to time-based
163 cleanup if the age argument is specified.
164 D
166 Similar to d, but in addition the contents of the directory will be
167 removed when --remove is used.
168 e
170 Adjust the mode and ownership of existing directories and remove
171 their contents based on age. Lines of this type accept shell-style
172 globs in place of normal path names. Contents of the directories
173 are subject to time-based cleanup if the age argument is specified.
174 If the age argument is "0", contents will be unconditionally
175 deleted every time systemd-tmpfiles --clean is run.
176 For this entry to be useful, at least one of the mode, user, group,
178 or age arguments must be specified, since otherwise this entry has
179 no effect. As an exception, an entry with no effect may be useful
180 when combined with !, see the examples.
181 v
183 Create a subvolume if the path does not exist yet, the file system
184 supports subvolumes (btrfs), and the system itself is installed
185 into a subvolume (specifically: the root directory / is itself a
186 subvolume). Otherwise, create a normal directory, in the same way
187 as d.
188 A subvolume created with this line type is not assigned to any
190 higher-level quota group. For that, use q or Q, which allow
191 creating simple quota group hierarchies, see below.
192 q
194 Create a subvolume or directory the same as v, but assign the
195 subvolume to the same higher-level quota groups as the parent. This
196 ensures that higher-level limits and accounting applied to the
197 parent subvolume also include the specified subvolume. On non-btrfs
198 file systems, this line type is identical to d.
199 If the subvolume already exists, no change to the quota hierarchy
201 is made, regardless of whether the subvolume is already attached to
202 a quota group or not. Also see Q below. See btrfs-qgroup(8) for
203 details about the btrfs quota group concept.
204 Q
206 Create the subvolume or directory the same as v, but assign the new
207 subvolume to a new leaf quota group. Instead of copying the
208 higher-level quota group assignments from the parent as is done
209 with q, the lowest quota group of the parent subvolume is
210 determined that is not the leaf quota group. Then, an
211 "intermediary" quota group is inserted that is one level below this
212 level, and shares the same ID part as the specified subvolume. If
213 no higher-level quota group exists for the parent subvolume, a new
214 quota group at level 255 sharing the same ID as the specified
215 subvolume is inserted instead. This new intermediary quota group is
216 then assigned to the parent subvolume's higher-level quota groups,
217 and the specified subvolume's leaf quota group is assigned to it.
218 Effectively, this has a similar effect as q, however introduces a
220 new higher-level quota group for the specified subvolume that may
221 be used to enforce limits and accounting to the specified subvolume
222 and children subvolume created within it. Thus, by creating
223 subvolumes only via q and Q, a concept of "subtree quotas" is
224 implemented. Each subvolume for which Q is set will get a "subtree"
225 quota group created, and all child subvolumes created within it
226 will be assigned to it. Each subvolume for which q is set will not
227 get such a "subtree" quota group, but it is ensured that they are
228 added to the same "subtree" quota group as their immediate parents.
229 It is recommended to use Q for subvolumes that typically contain
231 further subvolumes, and where it is desirable to have accounting
232 and quota limits on all child subvolumes together. Examples for Q
233 are typically /home/ or /var/lib/machines/. In contrast, q should
234 be used for subvolumes that either usually do not include further
235 subvolumes or where no accounting and quota limits are needed that
236 apply to all child subvolumes together. Examples for q are
237 typically /var/ or /var/tmp/.
238 As with q, Q has no effect on the quota group hierarchy if the
240 subvolume already exists, regardless of whether the subvolume
241 already belong to a quota group or not.
242 p, p+
244 Create a named pipe (FIFO) if it does not exist yet. If suffixed
245 with + and a file already exists where the pipe is to be created,
246 it will be removed and be replaced by the pipe.
247 L, L+
249 Create a symlink if it does not exist yet. If suffixed with + and a
250 file or directory already exists where the symlink is to be
251 created, it will be removed and be replaced by the symlink. If the
252 argument is omitted, symlinks to files with the same name residing
253 in the directory /usr/share/factory/ are created. Note that
254 permissions and ownership on symlinks are ignored.
255 c, c+
257 Create a character device node if it does not exist yet. If
258 suffixed with + and a file already exists where the device node is
259 to be created, it will be removed and be replaced by the device
260 node. It is recommended to suffix this entry with an exclamation
261 mark to only create static device nodes at boot, as udev will not
262 manage static device nodes that are created at runtime.
263 b, b+
265 Create a block device node if it does not exist yet. If suffixed
266 with + and a file already exists where the device node is to be
267 created, it will be removed and be replaced by the device node. It
268 is recommended to suffix this entry with an exclamation mark to
269 only create static device nodes at boot, as udev will not manage
270 static device nodes that are created at runtime.
271 C
273 Recursively copy a file or directory, if the destination files or
274 directories do not exist yet or the destination directory is empty.
275 Note that this command will not descend into subdirectories if the
276 destination directory already exists and is not empty. Instead, the
277 entire copy operation is skipped. If the argument is omitted, files
278 from the source directory /usr/share/factory/ with the same name
279 are copied. Does not follow symlinks. Contents of the directories
280 are subject to time-based cleanup if the age argument is specified.
281 x
283 Ignore a path during cleaning. Use this type to exclude paths from
284 clean-up as controlled with the Age parameter. Note that lines of
285 this type do not influence the effect of r or R lines. Lines of
286 this type accept shell-style globs in place of normal path names.
287 X
289 Ignore a path during cleaning. Use this type to exclude paths from
290 clean-up as controlled with the Age parameter. Unlike x, this
291 parameter will not exclude the content if path is a directory, but
292 only directory itself. Note that lines of this type do not
293 influence the effect of r or R lines. Lines of this type accept
294 shell-style globs in place of normal path names.
295 r
297 Remove a file or directory if it exists. This may not be used to
298 remove non-empty directories, use R for that. Lines of this type
299 accept shell-style globs in place of normal path names. Does not
300 follow symlinks.
301 R
303 Recursively remove a path and all its subdirectories (if it is a
304 directory). Lines of this type accept shell-style globs in place of
305 normal path names. Does not follow symlinks.
306 z
308 Adjust the access mode, user and group ownership, and restore the
309 SELinux security context of a file or directory, if it exists.
310 Lines of this type accept shell-style globs in place of normal path
311 names. Does not follow symlinks.
312 Z
314 Recursively set the access mode, user and group ownership, and
315 restore the SELinux security context of a file or directory if it
316 exists, as well as of its subdirectories and the files contained
317 therein (if applicable). Lines of this type accept shell-style
318 globs in place of normal path names. Does not follow symlinks.
319 t
321 Set extended attributes, see attr(5) for details. The argument
322 field should take one or more assignment expressions in the form
323 namespace.attribute=value, for examples see below. Lines of this
324 type accept shell-style globs in place of normal path names. This
325 can be useful for setting SMACK labels. Does not follow symlinks.
326 Please note that extended attributes settable with this line type
328 are a different concept from the Linux file attributes settable
329 with h/H, see below.
330 T
332 Same as t, but operates recursively.
333 h
335 Set Linux file/directory attributes. Lines of this type accept
336 shell-style globs in place of normal path names.
337 The format of the argument field is [+-=][aAcCdDeijPsStTu]. The
339 prefix + (the default one) causes the attributes to be added; -
340 causes the attributes to be removed; = causes the attributes to be
341 set exactly as the following letters. The letters "aAcCdDeijPsStTu"
342 select the new attributes for the files, see chattr(1) for further
343 information.
344 Passing only = as argument resets all the file attributes listed
346 above. It has to be pointed out that the = prefix limits itself to
347 the attributes corresponding to the letters listed here. All other
348 attributes will be left untouched. Does not follow symlinks.
349 Please note that the Linux file attributes settable with this line
351 type are a different concept from the extended attributes settable
352 with t/T, see above.
353 H
355 Sames as h, but operates recursively.
356 a, a+
358 Set POSIX ACLs (access control lists), see acl(5). If suffixed with
359 +, the specified entries will be added to the existing set.
360 systemd-tmpfiles will automatically add the required base entries
361 for user and group based on the access mode of the file, unless
362 base entries already exist or are explicitly specified. The mask
363 will be added if not specified explicitly or already present. Lines
364 of this type accept shell-style globs in place of normal path
365 names. This can be useful for allowing additional access to certain
366 files. Does not follow symlinks.
367 A, A+
369 Same as a and a+, but recursive. Does not follow symlinks.
370 Type Modifiers
372 If the exclamation mark ("!") is used, this line is only safe to
373 execute during boot, and can break a running system. Lines without the
374 exclamation mark are presumed to be safe to execute at any time, e.g.
375 on package upgrades. systemd-tmpfiles will take lines with an
376 exclamation mark only into consideration, if the --boot option is
377 given.
378 For example:
380 # Make sure these are created by default so that nobody else can
382 d /tmp/.X11-unix 1777 root root 10d
383 # Unlink the X11 lock files
385 r! /tmp/.X[0-9]*-lock
386 The second line in contrast to the first one would break a running
388 system, and will only be executed with --boot.
389 If the minus sign ("-") is used, this line failing to run successfully
391 during create (and only create) will not cause the execution of
392 systemd-tmpfiles to return an error.
393 For example:
395 # Modify sysfs but don't fail if we are in a container with a read-only /proc
397 w- /proc/sys/vm/swappiness - - - - 10
398 If the equals sign ("=") is used, the file types of existing objects in
400 the specified path are checked, and removed if they do not match. This
401 includes any implicitly created parent directories (which can be either
402 directories or directory symlinks). For example, if there is a FIFO in
403 place of one of the parent path components it will be replaced with a
404 directory.
405 If the tilde character ("~") is used, the argument (i.e. 6th) column is
407 Base64 decoded[1] before use. This modifier is only supported on line
408 types that can write file contents, i.e. f, f+, w, +. This is useful
409 for writing arbitrary binary data (including newlines and NUL bytes) to
410 files. Note that if this switch is used, the argument is not subject to
411 specifier expansion, neither before nor after Base64 decoding.
412 If the caret character ("^") is used, the argument (i.e. 6th) column
414 takes a service credential name to read the argument data from. See
415 System and Service Credentials[2] for details about the credentials
416 concept. This modifier is only supported on line types that can write
417 file contents, i.e. f, f+, w, w+. This is useful for writing arbitrary
418 files with contents sourced from elsewhere, including from VM or
419 container managers further up. If the specified credential is not set
420 for the systemd-tmpfiles service, the line is silently skipped. If "^"
421 and "~" are combined Base64 decoding is applied to the credential
422 contents.
423 Note that for all line types that result in creation of any kind of
425 file node (i.e. f/F, d/D/v/q/Q, p, L, c/b and C) leading directories
426 are implicitly created if needed, owned by root with an access mode of
427 0755. In order to create them with different modes or ownership make
428 sure to add appropriate d lines.
429 Path
431 The file system path specification supports simple specifier expansion,
432 see below. The path (after expansion) must be absolute.
433 Mode
435 The file access mode to use when creating this file or directory. If
436 omitted or when set to "-", the default is used: 0755 for directories,
437 0644 for all other file objects. For z, Z lines, if omitted or when set
438 to "-", the file access mode will not be modified. This parameter is
439 ignored for x, r, R, L, t, and a lines.
440 Optionally, if prefixed with "~", the access mode is masked based on
442 the already set access bits for existing file or directories: if the
443 existing file has all executable bits unset, all executable bits are
444 removed from the new access mode, too. Similarly, if all read bits are
445 removed from the old access mode, they will be removed from the new
446 access mode too, and if all write bits are removed, they will be
447 removed from the new access mode too. In addition, the sticky/SUID/SGID
448 bit is removed unless applied to a directory. This functionality is
449 particularly useful in conjunction with Z.
450 By default the access mode of listed inodes is set to the specified
452 mode regardless if it is created anew, or already existed. Optionally,
453 if prefixed with ":", the configured access mode is only applied when
454 creating new inodes, and if the inode the line refers to already
455 exists, its access mode is left in place unmodified.
456 User, Group
458 The user and group to use for this file or directory. This may either
459 be a numeric ID or a user/group name. If omitted or when set to "-",
460 the user and group of the user who invokes systemd-tmpfiles is used.
461 For z and Z lines, when omitted or when set to "-", the file ownership
462 will not be modified. These parameters are ignored for x, r, R, L, t,
463 and a lines.
464 This field should generally only reference system users/groups, i.e.
466 users/groups that are guaranteed to be resolvable during early boot. If
467 this field references users/groups that only become resolveable during
468 later boot (i.e. after NIS, LDAP or a similar networked directory
469 service become available), execution of the operations declared by the
470 line will likely fail. Also see Notes on Resolvability of User and
471 Group Names[3] for more information on requirements on system
472 user/group definitions.
473 By default the ownership of listed inodes is set to the specified
475 user/group regardless if it is created anew, or already existed.
476 Optionally, if prefixed with ":", the configured user/group information
477 is only applied when creating new inodes, and if the inode the line
478 refers to already exists, its user/group is left in place unmodified.
479 Age
481 The date field, when set, is used to decide what files to delete when
482 cleaning. If a file or directory is older than the current time minus
483 the age field, it is deleted. The field format is a series of integers
484 each followed by one of the following suffixes for the respective time
485 units: s, m or min, h, d, w, ms, and us, meaning seconds, minutes,
486 hours, days, weeks, milliseconds, and microseconds, respectively. Full
487 names of the time units can be used too.
488 If multiple integers and units are specified, the time values are
490 summed. If an integer is given without a unit, s is assumed.
491 When the age is set to zero, the files are cleaned unconditionally.
493 The age field only applies to lines starting with d, D, e, v, q, Q, C,
495 x and X. If omitted or set to "-", no automatic clean-up is done.
496 If the age field starts with a tilde character "~", clean-up is only
498 applied to files and directories one level inside the directory
499 specified, but not the files and directories immediately inside it.
500 The age of a file system entry is determined from its last modification
502 timestamp (mtime), its last access timestamp (atime), and (except for
503 directories) its last status change timestamp (ctime). By default, any
504 of these three (or two) values will prevent cleanup if it is more
505 recent than the current time minus the age field. To restrict the
506 deletion based on particular type of file timestamps, the age-by
507 argument can be used.
508 The age-by argument overrides the timestamp types to be used for the
510 age check. It can be specified by prefixing the age argument with a
511 sequence of characters to specify the timestamp types and a colon
512 (":"): "age-by...:cleanup-age". The argument can consist of a (A for
513 directories), b (B for directories), c (C for directories), or m (M for
514 directories). Those respectively indicate access, creation, last status
515 change, and last modification time of a file system entry. The
516 lower-case letter signifies that the given timestamp type should be
517 considered for files, while the upper-case letter signifies that the
518 given timestamp type should be considered for directories. See statx(2)
519 file timestamp fields for more details about timestamp types.
520 If not specified, the age-by field defaults to abcmABM, i.e. by default
522 all file timestamps are taken into consideration, with the exception of
523 the last status change timestamp (ctime) for directories. This is
524 because the aging logic itself will alter the ctime whenever it deletes
525 a file inside it. To ensure that running the aging logic does not feed
526 back into the next iteration of itself, ctime for directories is
527 ignored by default.
528 For example:
530 # Files created and modified, and directories accessed more than
532 # an hour ago in "/tmp/foo/bar", are subject to time-based cleanup.
533 d /tmp/foo/bar - - - bmA:1h -
534 Note that while the aging algorithm is run a 'shared' BSD file lock
536 (see flock(2)) is taken on each directory the algorithm descends into
537 (and each directory below that, and so on). If the aging algorithm
538 finds a lock is already taken on some directory, it (and everything
539 below it) is skipped. Applications may use this to temporarily exclude
540 certain directory subtrees from the aging algorithm: the applications
541 can take a BSD file lock themselves, and as long as they keep it aging
542 of the directory and everything below it is disabled.
543 Argument
545 For L lines determines the destination path of the symlink. For c and
546 b, determines the major/minor of the device node, with major and minor
547 formatted as integers, separated by ":", e.g. "1:3". For f, F, and w,
548 the argument may be used to specify a short string that is written to
549 the file, suffixed by a newline. For C, specifies the source file or
550 directory. For t and T, determines extended attributes to be set. For a
551 and A, determines ACL attributes to be set. For h and H, determines the
552 file attributes to set. Ignored for all other lines.
553 This field can contain specifiers, see below.
555
557 Specifiers can be used in the "path" and "argument" fields. An unknown
558 or unresolvable specifier is treated as invalid configuration. The
559 following expansions are understood:
560 Table 1. Specifiers available
562 ┌──────────┬─────────────────────┬────────────────────────┐
563 │Specifier │ Meaning │ Details │
564 ├──────────┼─────────────────────┼────────────────────────┤
565 │"%a" │ Architecture │ A short string │
566 │ │ │ identifying the │
567 │ │ │ architecture of the │
568 │ │ │ local system. A │
569 │ │ │ string such as x86, │
570 │ │ │ x86-64 or arm64. │
571 │ │ │ See the │
572 │ │ │ architectures │
573 │ │ │ defined for │
574 │ │ │ ConditionArchitecture= │
575 │ │ │ in systemd.unit(5) │
576 │ │ │ for a full list. │
577 ├──────────┼─────────────────────┼────────────────────────┤
578 │"%A" │ Operating system │ The operating system │
579 │ │ image version │ image version │
580 │ │ │ identifier of the │
581 │ │ │ running system, as │
582 │ │ │ read from the │
583 │ │ │ IMAGE_VERSION= field │
584 │ │ │ of /etc/os-release. If │
585 │ │ │ not set, resolves to │
586 │ │ │ an empty string. See │
587 │ │ │ os-release(5) for more │
588 │ │ │ information. │
589 ├──────────┼─────────────────────┼────────────────────────┤
590 │"%b" │ Boot ID │ The boot ID of the │
591 │ │ │ running system, │
592 │ │ │ formatted as string. │
593 │ │ │ See random(4) for more │
594 │ │ │ information. │
595 ├──────────┼─────────────────────┼────────────────────────┤
596 │"%B" │ Operating system │ The operating system │
597 │ │ build ID │ build identifier of │
598 │ │ │ the running system, as │
599 │ │ │ read from the │
600 │ │ │ BUILD_ID= field of │
601 │ │ │ /etc/os-release. If │
602 │ │ │ not set, resolves to │
603 │ │ │ an empty string. See │
604 │ │ │ os-release(5) for more │
605 │ │ │ information. │
606 ├──────────┼─────────────────────┼────────────────────────┤
607 │"%C" │ System or user │ In --user mode, this │
608 │ │ cache directory │ is the same as │
609 │ │ │ $XDG_CACHE_HOME, and │
610 │ │ │ /var/cache otherwise. │
611 ├──────────┼─────────────────────┼────────────────────────┤
612 │"%g" │ User group │ This is the name of │
613 │ │ │ the group running the │
614 │ │ │ command. In case of │
615 │ │ │ the system instance │
616 │ │ │ this resolves to │
617 │ │ │ "root". │
618 ├──────────┼─────────────────────┼────────────────────────┤
619 │"%G" │ User GID │ This is the numeric │
620 │ │ │ GID of the group │
621 │ │ │ running the command. │
622 │ │ │ In case of the system │
623 │ │ │ instance this resolves │
624 │ │ │ to 0. │
625 ├──────────┼─────────────────────┼────────────────────────┤
626 │"%h" │ User home directory │ This is the home │
627 │ │ │ directory of the user │
628 │ │ │ running the command. │
629 │ │ │ In case of the system │
630 │ │ │ instance this resolves │
631 │ │ │ to "/root". │
632 ├──────────┼─────────────────────┼────────────────────────┤
633 │"%H" │ Host name │ The hostname of the │
634 │ │ │ running system. │
635 ├──────────┼─────────────────────┼────────────────────────┤
636 │"%l" │ Short host name │ The hostname of the │
637 │ │ │ running system, │
638 │ │ │ truncated at the first │
639 │ │ │ dot to remove any │
640 │ │ │ domain component. │
641 ├──────────┼─────────────────────┼────────────────────────┤
642 │"%L" │ System or user log │ In --user mode, this │
643 │ │ directory │ is the same as │
644 │ │ │ $XDG_CONFIG_HOME with │
645 │ │ │ /log appended, and │
646 │ │ │ /var/log otherwise. │
647 ├──────────┼─────────────────────┼────────────────────────┤
648 │"%m" │ Machine ID │ The machine ID of the │
649 │ │ │ running system, │
650 │ │ │ formatted as string. │
651 │ │ │ See machine-id(5) for │
652 │ │ │ more information. │
653 ├──────────┼─────────────────────┼────────────────────────┤
654 │"%M" │ Operating system │ The operating system │
655 │ │ image identifier │ image identifier of │
656 │ │ │ the running system, as │
657 │ │ │ read from the │
658 │ │ │ IMAGE_ID= field of │
659 │ │ │ /etc/os-release. If │
660 │ │ │ not set, resolves to │
661 │ │ │ an empty string. See │
662 │ │ │ os-release(5) for more │
663 │ │ │ information. │
664 ├──────────┼─────────────────────┼────────────────────────┤
665 │"%o" │ Operating system ID │ The operating system │
666 │ │ │ identifier of the │
667 │ │ │ running system, as │
668 │ │ │ read from the ID= │
669 │ │ │ field of │
670 │ │ │ /etc/os-release. See │
671 │ │ │ os-release(5) for more │
672 │ │ │ information. │
673 ├──────────┼─────────────────────┼────────────────────────┤
674 │"%S" │ System or user │ In --user mode, this │
675 │ │ state directory │ is the same as │
676 │ │ │ $XDG_CONFIG_HOME, and │
677 │ │ │ /var/lib otherwise. │
678 ├──────────┼─────────────────────┼────────────────────────┤
679 │"%t" │ System or user │ In --user mode, this │
680 │ │ runtime directory │ is the same │
681 │ │ │ $XDG_RUNTIME_DIR, and │
682 │ │ │ /run/ otherwise. │
683 ├──────────┼─────────────────────┼────────────────────────┤
684 │"%T" │ Directory for │ This is either /tmp or │
685 │ │ temporary files │ the path "$TMPDIR", │
686 │ │ │ "$TEMP" or "$TMP" are │
687 │ │ │ set to. (Note that the │
688 │ │ │ directory may be │
689 │ │ │ specified without a │
690 │ │ │ trailing slash.) │
691 ├──────────┼─────────────────────┼────────────────────────┤
692 │"%u" │ User name │ This is the name of │
693 │ │ │ the user running the │
694 │ │ │ command. In case of │
695 │ │ │ the system instance │
696 │ │ │ this resolves to │
697 │ │ │ "root". │
698 ├──────────┼─────────────────────┼────────────────────────┤
699 │"%U" │ User UID │ This is the numeric │
700 │ │ │ UID of the user │
701 │ │ │ running the command. │
702 │ │ │ In case of the system │
703 │ │ │ instance this resolves │
704 │ │ │ to 0. │
705 ├──────────┼─────────────────────┼────────────────────────┤
706 │"%v" │ Kernel release │ Identical to uname -r │
707 │ │ │ output. │
708 ├──────────┼─────────────────────┼────────────────────────┤
709 │"%V" │ Directory for │ This is either │
710 │ │ larger and │ /var/tmp or the path │
711 │ │ persistent │ "$TMPDIR", "$TEMP" or │
712 │ │ temporary files │ "$TMP" are set to. │
713 │ │ │ (Note that the │
714 │ │ │ directory may be │
715 │ │ │ specified without a │
716 │ │ │ trailing slash.) │
717 ├──────────┼─────────────────────┼────────────────────────┤
718 │"%w" │ Operating system │ The operating system │
719 │ │ version ID │ version identifier of │
720 │ │ │ the running system, as │
721 │ │ │ read from the │
722 │ │ │ VERSION_ID= field of │
723 │ │ │ /etc/os-release. If │
724 │ │ │ not set, resolves to │
725 │ │ │ an empty string. See │
726 │ │ │ os-release(5) for more │
727 │ │ │ information. │
728 ├──────────┼─────────────────────┼────────────────────────┤
729 │"%W" │ Operating system │ The operating system │
730 │ │ variant ID │ variant identifier of │
731 │ │ │ the running system, as │
732 │ │ │ read from the │
733 │ │ │ VARIANT_ID= field of │
734 │ │ │ /etc/os-release. If │
735 │ │ │ not set, resolves to │
736 │ │ │ an empty string. See │
737 │ │ │ os-release(5) for more │
738 │ │ │ information. │
739 ├──────────┼─────────────────────┼────────────────────────┤
740 │"%%" │ Single percent sign │ Use "%%" in place of │
741 │ │ │ "%" to specify a │
742 │ │ │ single percent sign. │
743 └──────────┴─────────────────────┴────────────────────────┘
744
746 Example 1. Create directories with specific mode and ownership
747 screen(1), needs two directories created at boot with specific modes
749 and ownership:
750 # /usr/lib/tmpfiles.d/screen.conf
752 d /run/screens 1777 root screen 10d
753 d /run/uscreens 0755 root screen 10d12h
754 Contents of /run/screens and /run/uscreens will be cleaned up after 10
756 and 10½ days, respectively.
757 Example 2. Create a directory with a SMACK attribute
759 D /run/cups - - - -
761 t /run/cups - - - - security.SMACK64=printing user.attr-with-spaces="foo bar"
762 The directory will be owned by root and have default mode. Its contents
765 are not subject to time-based cleanup, but will be obliterated when
766 systemd-tmpfiles --remove runs.
767 Example 3. Create a directory and prevent its contents from cleanup
769 abrt(1), needs a directory created at boot with specific mode and
771 ownership and its content should be preserved from the automatic
772 cleanup applied to the contents of /var/tmp:
773 # /usr/lib/tmpfiles.d/tmp.conf
775 d /var/tmp 1777 root root 30d
776 # /usr/lib/tmpfiles.d/abrt.conf
778 d /var/tmp/abrt 0755 abrt abrt -
779 Example 4. Apply clean up during boot and based on time
781 # /usr/lib/tmpfiles.d/dnf.conf
783 r! /var/cache/dnf/*/*/download_lock.pid
784 r! /var/cache/dnf/*/*/metadata_lock.pid
785 r! /var/lib/dnf/rpmdb_lock.pid
786 e /var/cache/dnf/ - - - 30d
787 The lock files will be removed during boot. Any files and directories
789 in /var/cache/dnf/ will be removed after they have not been accessed in
790 30 days.
791 Example 5. Empty the contents of a cache directory on boot
793 # /usr/lib/tmpfiles.d/krb5rcache.conf
795 e! /var/cache/krb5rcache - - - 0
796 Any files and subdirectories in /var/cache/krb5rcache/ will be removed
798 on boot. The directory will not be created.
799 Example 6. Provision SSH public key access for root user via
801 Credentials in QEMU
802 -smbios type=11,value=io.systemd.credential.binary:tmpfiles.extra=$(echo "f~ /root/.ssh/authorized_keys 700 root root - $(ssh-add -L | base64 -w 0)" | base64 -w 0)
804 By passing this line to QEMU, the public key of the current user will
806 be encoded in base64, added to a tmpfiles.d line that tells
807 systemd-tmpfiles to decode it into /root/.ssh/authorized_keys, encode
808 that line itself in base64 and pass it as a Credential that will be
809 picked up by systemd from SMBIOS on boot.
810
812 /var/run/ is a deprecated symlink to /run/, and applications should use
813 the latter. systemd-tmpfiles will warn if /var/run/ is used.
814
816 systemd(1), systemd-tmpfiles(8), systemd-delta(1), systemd.exec(5),
817 attr(5), getfattr(1), setfattr(1), setfacl(1), getfacl(1), chattr(1),
818 btrfs-subvolume(8), btrfs-qgroup(8)
819
821 1. Base64 decoded
822 https://www.rfc-editor.org/rfc/rfc4648.html
823 2. System and Service Credentials
825 https://systemd.io/CREDENTIALS
826 3. Notes on Resolvability of User and Group Names
828 https://systemd.io/UIDS-GIDS/#notes-on-resolvability-of-user-and-group-names
829systemd 253 TMPFILES.D(5)