summaryrefslogtreecommitdiffabout
path: root/README
authorLars Hjemli <hjemli@gmail.com>2006-12-10 21:31:36 (UTC)
committer Lars Hjemli <hjemli@gmail.com>2006-12-10 21:31:36 (UTC)
commit25105d7ecaba474d4b7c364ebb586aac3dfc5abb (patch) (unidiff)
tree8beb08db1399b8efb8c7fbcd936044ae7fc232e6 /README
parent856c026e221d8ed82c5b75bc8da4bd65e89ea953 (diff)
downloadcgit-25105d7ecaba474d4b7c364ebb586aac3dfc5abb.zip
cgit-25105d7ecaba474d4b7c364ebb586aac3dfc5abb.tar.gz
cgit-25105d7ecaba474d4b7c364ebb586aac3dfc5abb.tar.bz2
Add caching infrastructure
This enables internal caching of page output. Page requests are split into four groups: 1) repo listing (front page) 2) repo summary 3) repo pages w/symbolic references in query string 4) repo pages w/constant sha1's in query string Each group has a TTL specified in minutes. When a page is requested, a cached filename is stat(2)'ed and st_mtime is compared to time(2). If TTL has expired (or the file didn't exist), the cached file is regenerated. When generating a cached file, locking is used to avoid parallell processing of the request. If multiple processes tries to aquire the same lock, the ones who fail to get the lock serves the (expired) cached file. If the cached file don't exist, the process instead calls sched_yield(2) before restarting the request processing. Signed-off-by: Lars Hjemli <hjemli@gmail.com>
Diffstat (limited to 'README') (more/less context) (ignore whitespace changes)
-rw-r--r--README54
1 files changed, 54 insertions, 0 deletions
diff --git a/README b/README
new file mode 100644
index 0000000..5917c37
--- a/dev/null
+++ b/README
@@ -0,0 +1,54 @@
1Cache algorithm
2===============
3
4Cgit normally returns cached pages when invoked. If there is no cache file, or
5the cache file has expired, it is regenerated. Finally, the cache file is
6printed on stdout.
7
8When it is decided that a cache file needs to be regenerated, an attempt is
9made to create a corresponding lockfile. If this fails, the process gives up
10and uses the expired cache file instead.
11
12When there is no cache file for a request, an attempt is made to create a
13corresponding lockfile. If this fails, the process calls sched_yield(2) before
14restarting the request handling.
15
16In pseudocode:
17
18 name = generate_cache_name(request);
19top:
20 if (!exists(name)) {
21 if (lock_cache(name)) {
22 generate_cache(request, name);
23 unlock_cache(name);
24 } else {
25 sched_yield();
26 goto top;
27 }
28 } else if (expired(name)) {
29 if (lock_cache(name)) {
30 generate_cache(request, name);
31 unlock_cache(name);
32 }
33 }
34 print_file(name);
35
36
37The following options can be set in /etc/cgitrc to control cache behaviour:
38 cache-root: root directory for cache files
39 cache-root-ttl: TTL for the repo listing page
40 cache-repo-ttl: TTL for any repos summary page
41 cache-dynamic-ttl: TTL for pages with symbolic references (not SHA1)
42 cache-static-ttl: TTL for pages with sha1 references
43
44TTL is specified in minutes, -1 meaning "infinite caching".
45
46
47Naming of cache files
48---------------------
49Repository listing: <cachedir>/index.html
50Repository summary: <cachedir>/<repo>/index.html
51Repository subpage: <cachedir>/<repo>/<page>/<querystring>.html
52
53The corresponding lock files have a ".lock" suffix.
54