3 # NOTE: additional options can be passed to git repack by specifying
4 # them after the project name, for example:
12 echo "Usage: gc.sh projname [extra-repack-args]" >&2
17 packopts
="--depth=50 --window=50 --window-memory=${var_window_memory:-1g}"
20 [ "$cfg_permission_control" != "Hooks" ] ||
umask 000
23 if _result
="$(kill -0 "$1" 2>&1)"; then
24 # process exists and we have permission to signal it
27 case "$_result" in *"not permitted"*)
28 # we do not have permission to signal the process
31 # process does not exist
36 # A .lock file should only exist for much less than a second.
37 # If we see a stale lock file (> 1h old), remove it and then,
38 # just in case, wait 30 seconds for any process whose .lock
39 # we might have just removed (it's racy) to finish doing what
40 # should take much less than a second to do.
41 _stalelock
="$(find "$1.lock
" -maxdepth 1 -mmin +60 -print 2>/dev/null || :)"
42 if [ -n "$_stalelock" ]; then
47 if (set -C; > "$1.lock") 2>/dev
/null
; then
52 [ "$_try" != "p" ] ||
sleep 1
54 # cannot create lock file
58 # return true if there's more than one objects/pack-<sha>.pack file or
59 # ANY sha-1 files in objects
61 _packs
=$
(find objects
/pack
-type f
-name "pack-$octet20.pack" -print |
wc -l)
62 if [ $_packs != 1 ] && [ $_packs != 0 ]; then
65 _objs
=$
(find objects
/$octet -type f
-name "$octet19" -print 2>/dev
/null |
wc -l)
69 # if the current directory is_gfi_mirror then repack all packs listed in gfi-packs
71 is_gfi_mirror ||
return 0
72 [ -s gfi-packs
] ||
return 0
73 while IFS
=': ' read -r _pack _junk
; do
74 if [ -s "$_pack" -a -s "${_pack%.pack}.idx" ]; then
75 git show-index
< "${_pack%.pack}.idx" | cut
-d ' ' -f 2
78 git pack-objects
$packopts --no-reuse-delta --delta-base-offset \
79 --non-empty $quiet objects
/pack
/packtmp | \
80 while read -r _newpack
; do
81 rm -f objects
/pack
/pack-
$_newpack.
*
82 ln -f objects
/pack
/packtmp-
$_newpack.pack objects
/pack
/pack-
$_newpack.pack
83 ln -f objects
/pack
/packtmp-
$_newpack.idx objects
/pack
/pack-
$_newpack.idx
84 rm -f objects
/pack
/packtmp-
$_newpack.
*
89 # HEADSHA="$(pack_is_complete /full/path/to/some.pack /full/path/to/packed-refs "$(cat HEAD)")"
91 # Must have a matching .idx file and a non-empty packed-refs file
92 [ -s "${1%.pack}.idx" ] ||
return 1
93 [ -s "$2" ] ||
return 1
99 "ref: refs/"?
*|
"ref:refs/"?
*|
"refs/"?
*)
100 _headmatch
="${3#ref:}"
101 _headmatch
="${_headmatch# }"
102 _headmatchpat
="$(echo "$_headmatch" | sed -e 's/\([.$]\)/\\\1/g')"
103 _headsha
="$(grep -e "^
$octet20 $_headmatchpat\$
" < "$2" | \
105 case "$_headsha" in $octet20) :;; *)
113 rm -rf pack_is_complete_test
114 mkdir pack_is_complete_test
115 mkdir pack_is_complete_test
/refs
116 mkdir pack_is_complete_test
/objects
117 mkdir pack_is_complete_test
/objects
/pack
118 echo "$_headsha" > pack_is_complete_test
/HEAD
119 ln -s "$1" pack_is_complete_test
/objects
/pack
/
120 ln -s "${1%.pack}.idx" pack_is_complete_test
/objects
/pack
/
121 ln -s "$2" pack_is_complete_test
/packed-refs
122 _count
="$(git --git-dir=pack_is_complete_test rev-list --count --all 2>/dev/null || :)"
123 rm -rf pack_is_complete_test
124 [ -n "$_count" ] ||
return 1
125 [ "$_count" -gt 0 ] 2>/dev
/null ||
return 1
131 cd "$cfg_reporoot/$proj.git"
133 trap 'if [ $? != 0 ]; then echo "gc failed dir: $PWD" >&2; fi' EXIT
137 # date -R is linux-only, POSIX equivalent is '+%a, %d %b %Y %T %z'
138 datefmt
='+%a, %d %b %Y %T %z'
140 if check_interval lastgc
$cfg_min_gc_interval; then
141 progress
"= [$proj] garbage check skip (last at $(config_get lastgc))"
144 if [ -e .nogc
]; then
145 progress
"x [$proj] garbage check disabled"
149 # Avoid unnecessary garbage collections:
150 # 1. If lastreceive is set and is older than lastgc
152 # 2. We are not a fork (! -s alternates) -OR- lastparentgc is older than lastgc
154 # If lastgc is NOT set or lastreceive is NOT set we MUST run gc
155 # If we are a fork and lastparentgc is NOT set we MUST run gc
157 # If the repo is dirty after removing any crud we MUST run gc
159 gcstart
="$(date "$datefmt")"
162 [ -s objects
/info
/alternates
] && isfork
=1
164 [ -n "$isfork" ] && lastparentgcsecs
="$(config_get_date_seconds lastparentgc || :)"
166 if lastreceivesecs
="$(config_get_date_seconds lastreceive)" && \
167 lastgcsecs
="$(config_get_date_seconds lastgc)" && \
168 [ $lastreceivesecs -lt $lastgcsecs ]; then
169 # We've run gc since we last received, so maybe we can skip,
170 # check if not fork or fork and lastparentgc < lastgc
171 if [ -n "$isfork" ]; then
172 if [ -n "$lastparentgcsecs" ] && \
173 [ $lastparentgcsecs -lt $lastgcsecs ]; then
174 # We've run gc since our parent ran gc so we can skip
178 # We don't have any alternates (we're not a forK) so we can skip
183 # be compatibile with gc.pid file from newer Git releases
188 if [ "$(createlock "$lockf")" ]; then
190 # 1) less than 12 hours old
191 # 2) contains two fields (pid hostname) NO trailing NL
192 # 3) the hostname is different OR the pid is still alive
193 # then we exit as another active process is holding the lock
194 if [ "$(find "$lockf" -maxdepth 1 -mmin -720 -print 2>/dev/null)" ]; then
197 read -r apid ahost ajunk
< "$lockf" ||
:
198 if [ "$apid" ] && [ "$ahost" ]; then
199 if [ "$ahost" != "$hn" ] || pidactive
"$apid"; then
205 echo >&2 "[$proj] unable to create gc.pid.lock file"
208 if [ -n "$active" ]; then
210 echo >&2 "[$proj] gc already running on machine '$ahost' pid '$apid'"
213 printf "%s %s" "$$" "$hn" > "$lockf.lock"
214 chmod 0664 "$lockf.lock"
215 mv -f "$lockf.lock" "$lockf"
217 # Remove any stale pack remnants that are more than an hour old.
218 # Stale pack fragments are defined as any pack-<sha1>.ext where .ext is NOT
219 # .pack AND the corresponding .pack DOES NOT exist. A bunch of stale
220 # pack-<sha1>.idx files without their corresponding .pack files are worthless
221 # and just waste space. Normally there shouldn't be any remnants but actually
222 # this can happen when things are interrupted at just the wrong time.
223 # Note that the objects/pack directory is created by git init and should
225 find objects
/pack
-maxdepth 1 -type f
-mmin +60 -name "pack-$octet20.?*" -print | \
226 sed -e 's/^objects\/pack\/pack-//; s/\..*$//' | LC_ALL
=C
sort -u | \
227 while read packsha
; do
228 [ ! -e "objects/pack/pack-$packsha.pack" ] ||
continue
229 rm -f "objects/pack/pack-$packsha".?
*
232 # Remove any stale pack .keep files that are more than 12 hours old.
233 # We don't do anything to create any permanent pack .keep files, so they must
234 # be remnants from some failed push or something. Removing the .keep will
235 # allow the pack to be properly repacked.
236 find objects
/pack
-maxdepth 1 -type f
-mmin +720 -name "pack-$octet20.keep" -print0 |
xargs -0 rm -f
238 # Remove any stale tmp_pack_* or tmp_idx_* or tmp_bitmap_* or packtmp-* files
239 # that are more than 12 hours old.
240 find objects
/pack
-maxdepth 1 -type f
-mmin +720 -name "tmp_pack_?*" -print0 |
xargs -0 rm -f
241 find objects
/pack
-maxdepth 1 -type f
-mmin +720 -name "tmp_idx_?*" -print0 |
xargs -0 rm -f
242 find objects
/pack
-maxdepth 1 -type f
-mmin +720 -name "tmp_bitmap_?*" -print0 |
xargs -0 rm -f
243 find objects
/pack
-maxdepth 1 -type f
-mmin +720 -name "packtmp-?*" -print0 |
xargs -0 rm -f
245 # Remove any stale shallow_* files that are more than 12 hours old.
246 # These can be left behind by Git >= 1.8.4.2 and < 2.0.0 when a client
247 # requests a shallow clone.
248 find .
-maxdepth 1 -type f
-mmin +720 -name "shallow_?*" -print0 |
xargs -0 rm -f
250 # Remove any stale *.temp files in the objects area that are more than 12 hours old.
251 # This can be stale sha1.temp, or stale *.pack.temp so we kill all stale *.temp.
252 find objects
-type f
-mmin +720 -name "*.temp" -print0 |
xargs -0 rm -f
254 # Remove any stale git-svn temp files that are more than 12 hours old.
255 # The git-svn process creates temp files with random 10 character names
256 # in the root of $GIT_DIR. Unfortunately they do not have a recognizable
257 # prefix, so we just have to kill any files with a 10-character name. We
258 # do this only for git-svn mirrors. All characters are chosen from
259 # [A-Za-z0-9_] so we can at least check that and fortunately the only
260 # collision is 'FETCH_HEAD' but that shouldn't matter.
261 # There may also be temp files with a Git_ prefix as well.
262 if is_svn_mirror
; then
263 _randchar
='[A-Za-z0-9_]'
264 _randchar2
="$_randchar$_randchar"
265 _randchar4
="$_randchar2$_randchar2"
266 _randchar10
="$_randchar4$_randchar4$_randchar2"
267 find .
-maxdepth 1 -type f
-mmin +720 -name "$_randchar10" -print0 |
xargs -0 rm -f
268 find .
-maxdepth 1 -type f
-mmin +720 -name "Git_*" -print0 |
xargs -0 rm -f
271 # Remove any stale fast_import_crash_<pid> files that are more than 3 days old.
272 if is_gfi_mirror
; then
273 find .
-maxdepth 1 -type f
-mmin +4320 -name "fast_import_crash_?*" -print0 |
xargs -0 rm -f
276 # Do not skip gc if the repo is dirty
277 if [ -n "$skipgc" ] && ! is_dirty
; then
278 progress
"= [$proj] garbage check nothing but crud removal to do (`date`)"
279 config_set lastgc
"$gcstart"
285 if [ -n "$isfork" ] && [ -z "$lastparentgcsecs" ]; then
286 # set lastparentgc and then update gcstart to be at least 1 second later
287 config_set lastparentgc
"$gcstart"
290 if [ -z "$lastreceivesecs" ]; then
291 # set lastreceive and then update gcstart to be at least 1 second later
292 config_set lastreceive
"$gcstart"
295 if [ -n "$bumptime" ]; then
297 gcstart
="$(date "$datefmt")"
300 progress
"+ [$proj] garbage check (`date`)"
302 # safe pruning: we put all our objects in all forks, then we can
303 # safely get rid of extra ones; repacks in forks will get rid of
304 # the redundant ones again then
306 if [ -d "../${forkdir##*/}" ]; then
307 # It is enough to copy objects just one level down and get_repo_list
308 # takes a regular expression (which is automatically prefixed with '^')
309 # so we can easily match forks exactly one level down from this project
310 get_repo_list
"$forkdir/[^/]*:" |
312 # Ignore forks that do not exist or are symbolic links
313 [ ! -L "$cfg_reporoot/$fork.git" -a -d "$cfg_reporoot/$fork.git" ] || \
315 # Or do not have a non-zero length alternates file
316 [ -s "$cfg_reporoot/$fork.git/objects/info/alternates" ] || \
318 # Match objects in parent project
319 for d
in objects
/?? objects
/pack
; do
320 [ "$d" != "objects/??" ] ||
continue
321 mkdir
-p "$cfg_reporoot/$fork.git/$d"
322 [ "$d" != "objects/pack" ] || \
323 [ "$(echo objects/pack/*)" != \
324 "objects/pack/*" ] || \
326 ln -f "$d"/* "$cfg_reporoot/$fork.git/$d" ||
:
328 # Update the fork's lastparentgc date (must be current, not $gcstart)
329 GIT_DIR
="$cfg_reporoot/$fork.git" git config \
330 gitweb.lastparentgc
"$(date "$datefmt")"
334 quiet
=; [ -n "$show_progress" ] || quiet
=-q
338 rm -f .gcfailed bundles
/*
339 rm -f objects
/pack
/pack-
*.bndl
340 # The git repack command may issue a 'disabling bitmap' warning for some
341 # repositories. This is perfectly normal and should be suppressed unless
342 # show_progress is set. Unfortunately that means we have to grep -v the
343 # output. And furthermore, since it's a translated message, we have to
344 # force the language to english to be sure we do it.
345 repackcmd
="git repack $packopts -a -d -l $quiet $@"
346 [ -n "$show_progress" ] || \
347 repackcmd
="{ LC_ALL=C $repackcmd 2>&1 || touch .gcfailed; } | grep -v 'disabling bitmap' || :"
349 [ ! -e .gcfailed
] ||
exit 1
350 allpacks
="$(echo objects/pack/pack-$octet20.pack)"
351 curhead
="$(cat HEAD)"
353 [ ! -e packed-refs
] || pkrf
=packed-refs
354 eval "reposizek=$(( $(echo 0 $(du -k $pkrf $allpacks 2>/dev/null | awk '{print $1}') | \
355 sed -e 's/ / + /g') ))"
356 # The git prune command does not take a -q or --quiet but started outputting
357 # 'Checking connectivity' progress messages in v1.7.9. However, we can
358 # suppress those by piping through cat as it only activates the progress
359 # messages when stderr is a tty.
361 [ -n "$show_progress" ] || \
362 prunecmd
="{ $prunecmd 2>&1 || touch .gcfailed; } | cat"
364 [ ! -e .gcfailed
] ||
exit 1
365 git update-server-info
367 # darcs:// mirrors have a xxx.log file that will grow endlessly
368 # if this is a mirror and the file exists, shorten it to 10000 lines
369 # also take this opportunity to optimize the darcs repo
370 if [ ! -e .nofetch
] && [ -n "$cfg_mirror" ]; then
371 url
="$(config_get baseurl || :)"
372 case "$url" in darcs
://*)
373 if [ -n "$cfg_mirror_darcs" ]; then
375 basedarcs
="$(basename "${url#darcs:/}")"
376 if [ -f "$basedarcs.log" ]; then
377 tail -n 10000 "$basedarcs.log" > "$basedarcs.log.$$"
378 mv -f "$basedarcs.log.$$" "$basedarcs.log"
380 if [ -d "$basedarcs.darcs" ]; then
382 cd "$basedarcs.darcs"
383 # Note that this does not optimize _darcs/inventories/ :(
391 # Create a matching .bndl header file for the all-in-one pack we just created
392 # but only if we're not a fork (otherwise the bundle would not be complete)
393 # and we are running at least Git version 1.7.2 (pack_is_complete always fails otherwise)
394 if [ ! -s objects
/info
/alternates
] && [ -n "$var_have_git_172" ]; then
395 # There should only be one pack in $allpacks but if there was a
396 # simultaneous push...
397 # The one we just created will have a .idx and will NOT have a .keep
400 for pk
in $allpacks; do
401 [ -s "$pk" ] ||
continue
403 [ -s "$pkbase.idx" ] ||
continue
404 [ ! -e "$pkbase.keep" ] ||
continue
405 if pkhead
="$(pack_is_complete "$PWD/$pk" "$PWD/packed-refs
" "$curhead")"; then
410 if [ -n "$pkfound" -a -n "$pkhead" ]; then
412 echo "# v2 git bundle"
413 sed -ne "/^$octet20 refs\/[^ $tab]*\$/ p" < packed-refs
417 bndletag
="$("$cfg_basedir/bin
/rangecgi
" --etag "$pkbase.bndl
" "$pkbase.pack
" || :)"
418 bndlsha
="$(printf '%s' "$bndletag" | git hash-object --stdin || :)"
419 if [ -n "$bndletag" ]; then
420 case "$bndlsha" in $octet20)
421 bndlshatrailer
="${bndlsha#????????}"
422 bndlshaprefix
="${bndlsha%$bndlshatrailer}"
423 bndlname
="$(TZ=UTC date +%Y%m%d_%H%M%S)-${bndlshaprefix:-0}"
424 [ -d bundles
] || mkdir bundles
425 echo "${pkbase#objects/pack/}.bndl" > "bundles/$bndlname"
426 echo "${pkbase#objects/pack/}.pack" >> "bundles/$bndlname"
432 # Record the size of this repo as the sum of its *.pack sizes as 1024-byte blocks
433 config_set_raw girocco.reposizek
"${reposizek:-0}"
435 # We use $gcstart here to avoid a race where a push occurs during the gc itself
436 # and the next future gc could be incorrectly skipped if we used the current
437 # timestamp here instead
438 config_set lastgc
"$gcstart"
441 progress
"- [$proj] garbage check (`date`)"