[Kde-pim] Akonadi(Next): Thoughts on caching
Milian Wolff
mail at milianw.de
Wed Jan 21 13:00:48 GMT 2015
On Wednesday 21 January 2015 11:46:07 Martin Steigerwald wrote:
> Hi!
>
> Considering:
>
> [Akonadi] [Bug 338402] File system cache is inneficient : too many file per
> directory
>
> Bug 332013 - NFS with NetApp FAS: please split payload files in file_db_data
> into several directories to avoid reaching maxdirsize limit on Ontap / WAFL
> filesystem
This should actually be trivial to fix. One just needs to add N levels of
directories inbetween. I.e. now the naming scheme is
$path/$id_$suffix
e.g.:
$path/304118_r0
$path/750546_r1
instead, it could do something like git does:
$path/$id[0]$id[1]/$id[2:]_$suffix
e.g.:
$path/30/4118_r0
$path/75/0546_r1
In my cache folder, that would produce the following:
#entries | prefix
11080 10
2032 11
567 12
1768 13
5444 14
1066 15
520 16
1823 17
2159 18
2722 19
2195 20
2502 21
2042 22
2271 23
2108 24
2274 25
2161 26
2317 27
2379 28
2688 29
1827 30
1303 31
978 32
1526 33
1702 34
1553 35
1759 36
1284 37
1903 38
1935 39
1540 40
1583 41
1521 42
2322 43
2548 44
2508 45
2303 46
2266 47
2793 48
2610 49
5142 50
3140 51
2586 52
2075 53
2513 54
1950 55
2310 56
2345 57
2458 58
1900 59
2139 60
1923 61
1984 62
1727 63
1722 64
1687 65
51 66
1425 67
1801 68
1667 69
1644 70
1569 71
1722 72
1118 73
2913 74
1983 75
1301 76
1683 77
1637 78
1974 79
1369 80
2143 81
1764 82
1818 83
1743 84
1973 85
1876 86
1901 87
2582 88
1923 89
2508 90
2383 91
3842 92
2564 93
2898 94
2194 95
2015 96
2199 97
2498 98
2549 99
of course, this could be further improved by using the last two integers of
the ID instead of the first (to prevent e.g. all files of ID 10... to be moved
into the 10 bucket, as seen above). It's still trivial to do the lookup by ID
then, definitely not much slower than before.
Bye
--
Milian Wolff
mail at milianw.de
http://milianw.de
More information about the kde-pim
mailing list