fsck: avoid reading every object twice
[git/git.git] / Documentation / git-fsck.txt
CommitLineData
df391b19
JH
1git-fsck(1)
2===========
3
4NAME
5----
6git-fsck - Verifies the connectivity and validity of the objects in the database
7
8
9SYNOPSIS
10--------
11[verse]
b1889c36 12'git fsck' [--tags] [--root] [--unreachable] [--cache] [--no-reflogs]
f29cd393 13 [--[no-]full] [--strict] [--verbose] [--lost-found] [<object>*]
df391b19
JH
14
15DESCRIPTION
16-----------
17Verifies the connectivity and validity of the objects in the database.
18
19OPTIONS
20-------
21<object>::
22 An object to treat as the head of an unreachability trace.
23+
0b444cdb 24If no objects are given, 'git fsck' defaults to using the
8d308b35
JK
25index file, all SHA1 references in .git/refs/*, and all reflogs (unless
26--no-reflogs is given) as heads.
df391b19
JH
27
28--unreachable::
43d532e6 29 Print out objects that exist but that aren't reachable from any
df391b19
JH
30 of the reference nodes.
31
32--root::
33 Report root nodes.
34
35--tags::
36 Report tags.
37
38--cache::
39 Consider any object recorded in the index also as a head node for
40 an unreachability trace.
41
566842f6
SP
42--no-reflogs::
43 Do not consider commits that are referenced only by an
44 entry in a reflog to be reachable. This option is meant
45 only to search for commits that used to be in a ref, but
46 now aren't, but are still in that corresponding reflog.
47
df391b19
JH
48--full::
49 Check not just objects in GIT_OBJECT_DIRECTORY
50 ($GIT_DIR/objects), but also the ones found in alternate
51 object pools listed in GIT_ALTERNATE_OBJECT_DIRECTORIES
52 or $GIT_DIR/objects/info/alternates,
53 and in packed git archives found in $GIT_DIR/objects/pack
54 and corresponding pack subdirectories in alternate
f29cd393
JH
55 object pools. This is now default; you can turn it off
56 with --no-full.
df391b19
JH
57
58--strict::
59 Enable more strict checking, namely to catch a file mode
60 recorded with g+w bit set, which was created by older
61 versions of git. Existing repositories, including the
62 Linux kernel, git itself, and sparse repository have old
63 objects that triggers this check, but it is recommended
64 to check new projects with this flag.
65
20f1eb6b
JS
66--verbose::
67 Be chatty.
68
68f6c019 69--lost-found::
16a7fcfe
JS
70 Write dangling objects into .git/lost-found/commit/ or
71 .git/lost-found/other/, depending on type. If the object is
72 a blob, the contents are written into the file, rather than
73 its object name.
68f6c019 74
df391b19
JH
75It tests SHA1 and general object sanity, and it does full tracking of
76the resulting reachability and everything else. It prints out any
77corruption it finds (missing or bad objects), and if you use the
78'--unreachable' flag it will also print out objects that exist but
43d532e6 79that aren't reachable from any of the specified head nodes.
df391b19
JH
80
81So for example
82
27c03aaf
MH
83 git fsck --unreachable HEAD \
84 $(git for-each-ref --format="%(objectname)" refs/heads)
df391b19
JH
85
86will do quite a _lot_ of verification on the tree. There are a few
87extra validity tests to be added (make sure that tree objects are
0b444cdb 88sorted properly etc), but on the whole if 'git fsck' is happy, you
df391b19
JH
89do have a valid tree.
90
91Any corrupt objects you will have to find in backups or other archives
2fd02c92 92(i.e., you can just remove them and do an 'rsync' with some other site in
df391b19
JH
93the hopes that somebody else has the object you have corrupted).
94
95Of course, "valid tree" doesn't mean that it wasn't generated by some
96evil person, and the end result might be crap. git is a revision
97tracking system, not a quality assurance system ;)
98
99Extracted Diagnostics
100---------------------
101
102expect dangling commits - potential heads - due to lack of head information::
103 You haven't specified any nodes as heads so it won't be
104 possible to differentiate between un-parented commits and
105 root nodes.
106
107missing sha1 directory '<dir>'::
108 The directory holding the sha1 objects is missing.
109
110unreachable <type> <object>::
111 The <type> object <object>, isn't actually referred to directly
112 or indirectly in any of the trees or commits seen. This can
113 mean that there's another root node that you're not specifying
114 or that the tree is corrupt. If you haven't missed a root node
115 then you might as well delete unreachable nodes since they
116 can't be used.
117
118missing <type> <object>::
119 The <type> object <object>, is referred to but isn't present in
120 the database.
121
122dangling <type> <object>::
123 The <type> object <object>, is present in the database but never
124 'directly' used. A dangling commit could be a root node.
125
df391b19
JH
126sha1 mismatch <object>::
127 The database has an object who's sha1 doesn't match the
128 database value.
129 This indicates a serious data integrity problem.
130
131Environment Variables
132---------------------
133
134GIT_OBJECT_DIRECTORY::
135 used to specify the object database root (usually $GIT_DIR/objects)
136
137GIT_INDEX_FILE::
138 used to specify the index file of the index
139
140GIT_ALTERNATE_OBJECT_DIRECTORIES::
141 used to specify additional object database roots (usually unset)
142
df391b19
JH
143GIT
144---
9e1f0a85 145Part of the linkgit:git[1] suite