1CACHE-CLEAN(1) NorduGrid Users Manual CACHE-CLEAN(1)
2
3
4
6 cache-clean - Administration tool for the A-REX cache.
7
8
10 cache-clean [-h] [-s] [-S] [-m NN -M NN] [-E N] [-D debug_level]
11 [-f space_command] [ -c <arex_config_file> | <dir1> [<dir2> [...]] ]
12
14 cache-clean is a tool for administrators of ARC server installations to
15 safely remove A-REX cache data and to provide an overview of the con‐
16 tents of the cache. It is used by the A-REX to automatically manage
17 cache contents.
18
19 There are two modes of operation - printing statistics and deleting
20 files. If -s is used, then statistics are printed on each cache. If -m
21 and -M are used then files in each cache are deleted if the space used
22 by the cache on the file system is more than that given by -M, in the
23 order of least recently accessed, until the space used by the cache is
24 equal to what is specified by -m. If -E is used, then all files ac‐
25 cessed less recently than the given time are deleted. -E can be used
26 in combination with -m and -M but deleting files using -E is carried
27 out first. If after this the cache used space is still more than that
28 given by -M then cleaning according to those options is performed.
29
30 If the cache is on a file system shared with other data then -S should
31 be specified so that the space used by the cache is calculated. Other‐
32 wise all the used space on the file system is assumed to be for the
33 cache. Using -S is slower so should only be used when the cache is
34 shared.
35
36 By default the "df" command is used to determine total and (if -S is
37 not specified) used space. If this command is not supported on the
38 cache file system then -f can be used to specify an alternate command.
39 The output of this command must be "total_bytes used_bytes", and so the
40 command would normally be a small script around the file system space
41 information tool. The cache directory is passed as the last argument to
42 this command.
43
44 Cache directories are given by dir1, dir2.. or taken from the config
45 file specified by -c or the ARC_CONFIG environment variable.
46
47 -h - print short help
48
49 -s - print cache statistics, without deleting anything. The output dis‐
50 plays for each cache the number of deletable (and locked) files, the
51 total size of these files, the percentage usage of the file system in
52 which the cache is stored, and a histogram of access times of the files
53 in the cache.
54
55 -S - Calculate the size of the cache instead of taking used space on
56 the file system. This should only be used when the cache file system is
57 shared with other data.
58
59 -M - the maximum used space (as % of the file system) at which to start
60 cleaning
61
62 -m - the minimum used space (as % of the file system) at which to stop
63 cleaning
64
65 -E - files accessed less recently than the given time period will be
66 deleted. Example values of this option are 1800, 90s, 24h, 30d. The de‐
67 fault when no suffix is given is seconds.
68
69 -f - alternative command to "df" for obtaining the file system total
70 and used space. The output of this command must be "total_bytes
71 used_bytes". The cache directory is passed as the last argument to this
72 command.
73
74 -D - debug level. Possible values are FATAL, ERROR, WARNING, INFO, VER‐
75 BOSE or DEBUG. Default level is INFO.
76
77 -c - path to an A-REX config file, xml or ini format
78
79 This tool is run periodically by the A-REX to keep the size of each
80 cache within the limits specified in the configuration file. Therefore
81 cleaning should not be performed manually, unless the cache size needs
82 to be reduced temporarily. For performance reasons it may however be
83 desirable to run cache-clean independently on the machine hosting the
84 cache file system, if this is different from the A-REX host. The most
85 useful function for administrators is to give an overview of the con‐
86 tents of the cache, using the -s option.
87
88 Within each cache directory specified in the configuration file, there
89 is a subdirectory for data (data/) and one for per-job hard links
90 (joblinks/). See the A-REX Administration Guide for more details.
91 cache-clean should only operate on the data subdirectory, therefore
92 when giving dir arguments they should be the top level cache directory.
93 cache-clean will then automatically only look at files within the data
94 directory.
95
96
97
99 cache-clean -m20 -M30 -E30d -D VERBOSE -c /etc/arc.conf
100
101 Cache directories are taken from the configuration file /etc/arc.conf
102 and all cache files accessed more than 30 days ago are deleted. Then if
103 the used space in a cache is above 30%, data is deleted until the used
104 space reaches 20%. Verbose debug output is enabled so information is
105 output on each file that is deleted.
106
107
109 APACHE LICENSE Version 2.0
110
111
113 ARC software is developed by the NorduGrid Collaboration
114 (http://www.nordugrid.org), please consult the AUTHORS file distributed
115 with ARC. Please report bugs and feature requests to
116 http://bugzilla.nordugrid.org
117
118
119
120
121NorduGrid ARC 6.13.0 2021-09-15 CACHE-CLEAN(1)