Alessandro Vermeulen

Alessandro Vermeulen's blog about languages, programming, Computer Science, photography and the Web.

Caching Hackage

| Comments

On several occasions I noticed that when performing a cabal update that the index was being downloaded at the rate of plus min 300 KB/s. Finally I got around to do something about this. I’ve set up a caching server located in Utrecht, The Netherlands. It is a caching proxy for the hackage repository. If you want to use it, add the following to your ~/.cabal/config file. (Or equivalent on Windows.)

1

remote-repo: cache:http://spockz.nl:12080/packages/archive

Be sure to comment out the already existing remote-repo. Otherwise, cabal will download both indexes and merge them, and we don’t want this.

The funny bit

Apparently this only helps if your machine is fast enough to process the index (untarring and all extra administration cabal performs).

Plainly getting the file from the cache:

1
2
3
4

> wget http://spockz.nl:12080/packages/archive/00-index.tar.gz
Saving to: 00-index.tar.gz
100%[======================================>] 3.394.821   10,9M/s   in 0,3s
2011-11-13 00:12:35 (10,9 MB/s) - 00-index.tar.gz saved [3394821/3394821]

And running cabal update with my cache as source:

1

cabal update -v3  7,50s user 0,21s system 99% cpu 7,736 total

And then finally, with the original repository:

1

cabal update -v3  7,57s user 0,25s system 28% cpu 27,372 total

So here we see that the user time is roughly the same but you spent almost three times more seconds waiting for your coffee to get cold. Any further speed improvements for cabal update will probably require optimalisation of the code.

The caching server

I’m using Varnish to cache the request to hackage. And here is my config file. Please shoot if you see any improvements.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106

# This is a basic VCL configuration file for varnish.  See the vcl(7)
# man page for details on VCL syntax and semantics.

backend backend_0 {
	.host = "hackage.haskell.org";
	.port = "80";
	.connect_timeout = 0.4s;
	.first_byte_timeout = 300s;
	.between_bytes_timeout = 60s;
}


acl purge {
  "localhost";
	"hackage.haskell.org";
}

sub vcl_recv {
    set req.grace = 120s;
    set req.backend = backend_0;

		set req.http.host = "hackage.haskell.org";

    if (req.request == "PURGE") {
        if (!client.ip ~ purge) {
            error 405 "Not allowed.";
        }
        return(lookup);
    }

    if (req.request != "GET" &&
        req.request != "HEAD" &&
        req.request != "PUT" &&
        req.request != "POST" &&
        req.request != "TRACE" &&
        req.request != "OPTIONS" &&
        req.request != "DELETE") {
        /* Non-RFC2616 or CONNECT which is weird. */
        return(pipe);
    }

    if (req.request != "GET" && req.request != "HEAD") {
        /* We only deal with GET and HEAD by default */
        return(pass);
    }

    if (req.http.If-None-Match) {
        return(pass);
    }

    if (req.url ~ "createObject") {
        return(pass);
    }

    remove req.http.Accept-Encoding;

    return(lookup);
}

sub vcl_pipe {
    # This is not necessary if you do not do any request rewriting.
    set req.http.connection = "close";

}

sub vcl_hit {

    if (req.request == "PURGE") {
        purge_url(req.url);
        error 200 "Purged";
    }

    if (!obj.cacheable) {
        return(pass);
    }
}

sub vcl_miss {

    if (req.request == "PURGE") {
        error 404 "Not in cache";
    }

}

sub vcl_fetch {
    set beresp.grace = 120s;

    if (!beresp.cacheable) {
        return(pass);
    }
    if (beresp.http.Set-Cookie) {
        return(pass);
    }
    if (beresp.http.Cache-Control ~ "(private|no-cache|no-store)") {
        return(pass);
    }
    if (beresp.http.Authorization && !beresp.http.Cache-Control ~ "public") {
        return(pass);
    }

}

sub vcl_deliver {

}

Comments