1GIT-FSCK(1) Git Manual GIT-FSCK(1)
2
6 git-fsck - Verifies the connectivity and validity of the objects in the
7 database
8
10 git fsck [--tags] [--root] [--unreachable] [--cache] [--no-reflogs]
11 [--[no-]full] [--strict] [--verbose] [--lost-found]
12 [--[no-]dangling] [--[no-]progress] [--connectivity-only]
13 [--[no-]name-objects] [<object>*]
14
16 Verifies the connectivity and validity of the objects in the database.
17
19 <object>
20 An object to treat as the head of an unreachability trace.
21 If no objects are given, git fsck defaults to using the index file,
23 all SHA-1 references in refs namespace, and all reflogs (unless
24 --no-reflogs is given) as heads.
25 --unreachable
27 Print out objects that exist but that aren’t reachable from any of
28 the reference nodes.
29 --[no-]dangling
31 Print objects that exist but that are never directly used
32 (default). --no-dangling can be used to omit this information from
33 the output.
34 --root
36 Report root nodes.
37 --tags
39 Report tags.
40 --cache
42 Consider any object recorded in the index also as a head node for
43 an unreachability trace.
44 --no-reflogs
46 Do not consider commits that are referenced only by an entry in a
47 reflog to be reachable. This option is meant only to search for
48 commits that used to be in a ref, but now aren’t, but are still in
49 that corresponding reflog.
50 --full
52 Check not just objects in GIT_OBJECT_DIRECTORY ($GIT_DIR/objects),
53 but also the ones found in alternate object pools listed in
54 GIT_ALTERNATE_OBJECT_DIRECTORIES or
55 $GIT_DIR/objects/info/alternates, and in packed Git archives found
56 in $GIT_DIR/objects/pack and corresponding pack subdirectories in
57 alternate object pools. This is now default; you can turn it off
58 with --no-full.
59 --connectivity-only
61 Check only the connectivity of reachable objects, making sure that
62 any objects referenced by a reachable tag, commit, or tree is
63 present. This speeds up the operation by avoiding reading blobs
64 entirely (though it does still check that referenced blobs exist).
65 This will detect corruption in commits and trees, but not do any
66 semantic checks (e.g., for format errors). Corruption in blob
67 objects will not be detected at all.
68 Unreachable tags, commits, and trees will also be accessed to find
70 the tips of dangling segments of history. Use --no-dangling if you
71 don’t care about this output and want to speed it up further.
72 --strict
74 Enable more strict checking, namely to catch a file mode recorded
75 with g+w bit set, which was created by older versions of Git.
76 Existing repositories, including the Linux kernel, Git itself, and
77 sparse repository have old objects that triggers this check, but it
78 is recommended to check new projects with this flag.
79 --verbose
81 Be chatty.
82 --lost-found
84 Write dangling objects into .git/lost-found/commit/ or
85 .git/lost-found/other/, depending on type. If the object is a blob,
86 the contents are written into the file, rather than its object
87 name.
88 --name-objects
90 When displaying names of reachable objects, in addition to the
91 SHA-1 also display a name that describes how they are reachable,
92 compatible with git-rev-parse(1), e.g.
93 HEAD@{1234567890}~25^2:src/.
94 --[no-]progress
96 Progress status is reported on the standard error stream by default
97 when it is attached to a terminal, unless --no-progress or
98 --verbose is specified. --progress forces progress status even if
99 the standard error stream is not directed to a terminal.
100
102 fsck.<msg-id>
103 During fsck git may find issues with legacy data which wouldn’t be
104 generated by current versions of git, and which wouldn’t be sent
105 over the wire if transfer.fsckObjects was set. This feature is
106 intended to support working with legacy repositories containing
107 such data.
108 Setting fsck.<msg-id> will be picked up by git-fsck(1), but to
110 accept pushes of such data set receive.fsck.<msg-id> instead, or to
111 clone or fetch it set fetch.fsck.<msg-id>.
112 The rest of the documentation discusses fsck.* for brevity, but
114 the same applies for the corresponding receive.fsck.* and
115 fetch.<msg-id>.*. variables.
116 Unlike variables like color.ui and core.editor the
118 receive.fsck.<msg-id> and fetch.fsck.<msg-id> variables will not
119 fall back on the fsck.<msg-id> configuration if they aren’t set. To
120 uniformly configure the same fsck settings in different
121 circumstances all three of them they must all set to the same
122 values.
123 When fsck.<msg-id> is set, errors can be switched to warnings and
125 vice versa by configuring the fsck.<msg-id> setting where the
126 <msg-id> is the fsck message ID and the value is one of error, warn
127 or ignore. For convenience, fsck prefixes the error/warning with
128 the message ID, e.g. "missingEmail: invalid author/committer line -
129 missing email" means that setting fsck.missingEmail = ignore will
130 hide that issue.
131 In general, it is better to enumerate existing objects with
133 problems with fsck.skipList, instead of listing the kind of
134 breakages these problematic objects share to be ignored, as doing
135 the latter will allow new instances of the same breakages go
136 unnoticed.
137 Setting an unknown fsck.<msg-id> value will cause fsck to die, but
139 doing the same for receive.fsck.<msg-id> and fetch.fsck.<msg-id>
140 will only cause git to warn.
141 fsck.skipList
143 The path to a list of object names (i.e. one unabbreviated SHA-1
144 per line) that are known to be broken in a non-fatal way and should
145 be ignored. On versions of Git 2.20 and later comments (#), empty
146 lines, and any leading and trailing whitespace is ignored.
147 Everything but a SHA-1 per line will error out on older versions.
148 This feature is useful when an established project should be
150 accepted despite early commits containing errors that can be safely
151 ignored such as invalid committer email addresses. Note: corrupt
152 objects cannot be skipped with this setting.
153 Like fsck.<msg-id> this variable has corresponding
155 receive.fsck.skipList and fetch.fsck.skipList variants.
156 Unlike variables like color.ui and core.editor the
158 receive.fsck.skipList and fetch.fsck.skipList variables will not
159 fall back on the fsck.skipList configuration if they aren’t set. To
160 uniformly configure the same fsck settings in different
161 circumstances all three of them they must all set to the same
162 values.
163 Older versions of Git (before 2.20) documented that the object
165 names list should be sorted. This was never a requirement, the
166 object names could appear in any order, but when reading the list
167 we tracked whether the list was sorted for the purposes of an
168 internal binary search implementation, which could save itself some
169 work with an already sorted list. Unless you had a humongous list
170 there was no reason to go out of your way to pre-sort the list.
171 After Git version 2.20 a hash implementation is used instead, so
172 there’s now no reason to pre-sort the list.
173
175 git-fsck tests SHA-1 and general object sanity, and it does full
176 tracking of the resulting reachability and everything else. It prints
177 out any corruption it finds (missing or bad objects), and if you use
178 the --unreachable flag it will also print out objects that exist but
179 that aren’t reachable from any of the specified head nodes (or the
180 default set, as mentioned above).
181 Any corrupt objects you will have to find in backups or other archives
183 (i.e., you can just remove them and do an rsync with some other site in
184 the hopes that somebody else has the object you have corrupted).
185 If core.commitGraph is true, the commit-graph file will also be
187 inspected using git commit-graph verify. See git-commit-graph(1).
188
190 expect dangling commits - potential heads - due to lack of head
191 information
192 You haven’t specified any nodes as heads so it won’t be possible to
193 differentiate between un-parented commits and root nodes.
194 missing sha1 directory <dir>
196 The directory holding the sha1 objects is missing.
197 unreachable <type> <object>
199 The <type> object <object>, isn’t actually referred to directly or
200 indirectly in any of the trees or commits seen. This can mean that
201 there’s another root node that you’re not specifying or that the
202 tree is corrupt. If you haven’t missed a root node then you might
203 as well delete unreachable nodes since they can’t be used.
204 missing <type> <object>
206 The <type> object <object>, is referred to but isn’t present in the
207 database.
208 dangling <type> <object>
210 The <type> object <object>, is present in the database but never
211 directly used. A dangling commit could be a root node.
212 hash mismatch <object>
214 The database has an object whose hash doesn’t match the object
215 database value. This indicates a serious data integrity problem.
216
218 GIT_OBJECT_DIRECTORY
219 used to specify the object database root (usually $GIT_DIR/objects)
220 GIT_INDEX_FILE
222 used to specify the index file of the index
223 GIT_ALTERNATE_OBJECT_DIRECTORIES
225 used to specify additional object database roots (usually unset)
226
228 Part of the git(1) suite
229Git 2.26.2 2020-04-20 GIT-FSCK(1)