1 _______________________
2 BUILDING THE FSQA SUITE
3 _______________________
6 - cd into the xfstests directory
7 - install prerequisite packages
8 For example, for Ubuntu:
9 "sudo apt-get install xfslibs-dev uuid-dev libtool-bin e2fsprogs
10 automake gcc libuuid1 quota attr libattr1-dev make
11 libacl1-dev libaio-dev xfsprogs libgdbm-dev gawk fio dbench
15 - create fsgqa test user ("sudo useradd fsgqa")
16 - create 123456-fsgqa test user ("sudo useradd 123456-fsgqa")
19 - cd into the xfstests directory
20 - set the ROOT and TOOLROOT env variables for IRIX appropriately
23 ______________________
25 ______________________
27 Preparing system for tests (IRIX and Linux):
29 - compile XFS into your kernel or load XFS modules
30 - install user tools including mkfs.xfs, xfs_db & xfs_bmap
31 - If you wish to run the udf components of the suite install
32 mkfs_udf and udf_db for IRIX and mkudffs for Linux. Also download and
33 build the Philips UDF Verification Software from
34 http://www.extra.research.philips.com/udf/, then copy the udf_test
35 binary to xfstests/src/. If you wish to disable UDF verification test
36 set the environment variable DISABLE_UDF_TEST to 1.
39 - create one or two partitions to use for testing
41 - format as XFS, mount & optionally populate with
43 - one SCRATCH partition (optional)
44 - leave empty and expect this partition to be clobbered
45 by some tests. If this is not provided, many tests will
47 (SCRATCH and TEST must be two DIFFERENT partitions)
49 - for btrfs only: some btrfs test cases will need 3 or more independent
50 SCRATCH disks which should be set using SCRATCH_DEV_POOL (for eg:
51 SCRATCH_DEV_POOL="/dev/sda /dev/sdb /dev/sdc") with which
52 SCRATCH_DEV should be unused by the tester, and for the legacy
53 support SCRATCH_DEV will be set to the first disk of the
54 SCRATCH_DEV_POOL by xfstests script.
56 - setup your environment
57 - setenv TEST_DEV "device containing TEST PARTITION"
58 - setenv TEST_DIR "mount point of TEST PARTITION"
60 - setenv SCRATCH_DEV "device containing SCRATCH PARTITION" OR
61 (btrfs only) setenv SCRATCH_DEV_POOL "to 3 or more SCRATCH disks for
62 testing btrfs raid concepts"
63 - setenv SCRATCH_MNT "mount point for SCRATCH PARTITION"
64 - setenv TAPE_DEV "tape device for testing xfsdump"
65 - setenv RMT_TAPE_DEV "remote tape device for testing xfsdump"
66 - setenv RMT_IRIXTAPE_DEV "remote IRIX tape device for testing xfsdump"
67 - setenv SCRATCH_LOGDEV "device for scratch-fs external log"
68 - setenv SCRATCH_RTDEV "device for scratch-fs realtime data"
69 - setenv TEST_LOGDEV "device for test-fs external log"
70 - setenv TEST_RTDEV "device for test-fs realtime data"
71 - if TEST_LOGDEV and/or TEST_RTDEV, these will always be used.
72 - if SCRATCH_LOGDEV and/or SCRATCH_RTDEV, the USE_EXTERNAL
73 environment variable set to "yes" will enable their use.
74 - setenv DIFF_LENGTH "number of diff lines to print from a failed test",
75 by default 10, set to 0 to print the full diff
76 - setenv FSTYP "the filesystem you want to test", the filesystem
77 type is devised from the TEST_DEV device, but you may want to
78 override it; if unset, the default is 'xfs'
79 - setenv FSSTRESS_AVOID and/or FSX_AVOID, which contain options
80 added to the end of fsstresss and fsx invocations, respectively,
81 in case you wish to exclude certain operational modes from these
83 - set TEST_XFS_REPAIR_REBUILD=1 to have _check_xfs_filesystem
84 run xfs_repair -n to check the filesystem; xfs_repair to rebuild
85 metadata indexes; and xfs_repair -n (a third time) to check the
86 results of the rebuilding.
88 - or add a case to the switch in common/config assigning
89 these variables based on the hostname of your test
91 - or add these variables to a file called local.config and keep that
92 file in your workarea.
94 - if testing xfsdump, make sure the tape devices have a
95 tape which can be overwritten.
97 - make sure $TEST_DEV is a mounted XFS partition
98 - make sure that $SCRATCH_DEV or $SCRATCH_DEV_POOL contains nothing useful
103 - By default the tests suite will run xfs tests:
104 - ./check '*/001' '*/002' '*/003'
106 - You can explicitly specify NFS/CIFS/UDF, otherwise the filesystem type will
107 be autodetected from $TEST_DEV:
108 ./check -nfs [test(s)]
109 - Groups of tests maybe ran by: ./check -g [group(s)]
110 See the 'group' file for details on groups
111 - for udf tests: ./check -udf [test(s)]
112 Running all the udf tests: ./check -udf -g udf
113 - for running nfs tests: ./check -nfs [test(s)]
114 - for running cifs/smb3 tests: ./check -cifs [test(s)]
115 - To randomize test order: ./check -r [test(s)]
118 The check script tests the return value of each script, and
119 compares the output against the expected output. If the output
120 is not as expected, a diff will be output and an .out.bad file
121 will be produced for the failing test.
123 Unexpected console messages, crashes and hangs may be considered
124 to be failures but are not necessarily detected by the QA system.
126 __________________________
127 ADDING TO THE FSQA SUITE
128 __________________________
131 Creating new tests scripts:
133 Use the "new" script.
135 Test script environment:
137 When developing a new test script keep the following things in
138 mind. All of the environment variables and shell procedures are
139 available to the script once the "common/rc" file has been
142 1. The tests are run from an arbitrary directory. If you want to
143 do operations on an XFS filesystem (good idea, eh?), then do
144 one of the following:
146 (a) Create directories and files at will in the directory
147 $TEST_DIR ... this is within an XFS filesystem and world
148 writeable. You should cleanup when your test is done,
149 e.g. use a _cleanup shell procedure in the trap ... see
150 001 for an example. If you need to know, the $TEST_DIR
151 directory is within the filesystem on the block device
154 (b) mkfs a new XFS filesystem on $SCRATCH_DEV, and mount this
155 on $SCRATCH_MNT. Call the the _require_scratch function
156 on startup if you require use of the scratch partition.
157 _require_scratch does some checks on $SCRATCH_DEV &
158 $SCRATCH_MNT and makes sure they're unmounted. You should
159 cleanup when your test is done, and in particular unmount
161 Tests can make use of $SCRATCH_LOGDEV and $SCRATCH_RTDEV
162 for testing external log and realtime volumes - however,
163 these tests need to simply "pass" (e.g. cat $seq.out; exit
164 - or default to an internal log) in the common case where
165 these variables are not set.
167 2. You can safely create temporary files that are not part of the
168 filesystem tests (e.g. to catch output, prepare lists of things
169 to do, etc.) in files named $tmp.<anything>. The standard test
170 script framework created by "new" will initialize $tmp and
173 3. By default, tests are run as the same uid as the person
174 executing the control script "check" that runs the test scripts.
176 4. Some other useful shell procedures:
178 _get_fqdn - echo the host's fully qualified
181 _get_pids_by_name - one argument is a process name, and
182 return all of the matching pids on
185 _within_tolerance - fancy numerical "close enough is good
186 enough" filter for deterministic
187 output ... see comments in
188 common/filter for an explanation
190 _filter_date - turn ctime(3) format dates into the
191 string DATE for deterministic
194 _cat_passwd, - dump the content of the password
195 _cat_group or group file (both the local file
196 and the content of the NIS database
197 if it is likely to be present)
199 5. General recommendations, usage conventions, etc.:
200 - When the content of the password or group file is
201 required, get it using the _cat_passwd and _cat_group
202 functions, to ensure NIS information is included if NIS
204 - When calling getfacl in a test, pass the "-n" argument so
205 that numeric rather than symbolic identifiers are used in
207 - When creating a new test, it is possible to enter a custom name
208 for the file. Filenames are in form NNN-custom-name, where NNN
209 is automatically added by the ./new script as an unique ID,
210 and "custom-name" is the optional string entered into a prompt
211 in the ./new script. It can contain only alphanumeric characters
212 and dash. Note the "NNN-" part is added automatically.
216 Each test script has a name, e.g. 007, and an associated
217 verified output, e.g. 007.out.
219 It is important that the verified output is deterministic, and
220 part of the job of the test script is to filter the output to
221 make this so. Examples of the sort of things that need filtering:
228 - variable directory contents
229 - imprecise numbers, especially sizes and times
233 The script "check" may be used to run one or more tests.
235 Test number $seq is deemed to "pass" when:
236 (a) no "core" file is created,
237 (b) the file $seq.notrun is not created,
238 (c) the exit status is 0, and
239 (d) the output matches the verified output.
241 In the "not run" case (b), the $seq.notrun file should contain a
242 short one-line summary of why the test was not run. The standard
243 output is not checked, so this can be used for a more verbose
244 explanation and to provide feedback when the QA test is run
248 To force a non-zero exit status use:
254 won't have the desired effect because of the way the exit trap
257 The recent pass/fail history is maintained in the file "check.log".
258 The elapsed time for the most recent pass for each test is kept