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 e2fsprogs
10 automake gcc libuuid1 quota attr libattr1-dev
11 libacl1-dev libaio-dev xfsprogs libgdbm-dev gawk fio dbench"
14 - create fsgqa test user ("sudo useradd fsgqa")
17 - cd into the xfstests directory
18 - set the ROOT and TOOLROOT env variables for IRIX appropriately
21 ______________________
23 ______________________
25 Preparing system for tests (IRIX and Linux):
27 - compile XFS into your kernel or load XFS modules
28 - install user tools including mkfs.xfs, xfs_db & xfs_bmap
29 - If you wish to run the udf components of the suite install
30 mkfs_udf and udf_db for IRIX and mkudffs for Linux. Also download and
31 build the Philips UDF Verification Software from
32 http://www.extra.research.philips.com/udf/, then copy the udf_test
33 binary to xfstests/src/. If you wish to disable UDF verification test
34 set the environment variable DISABLE_UDF_TEST to 1.
37 - create one or two partitions to use for testing
39 - format as XFS, mount & optionally populate with
41 - one SCRATCH partition (optional)
42 - leave empty and expect this partition to be clobbered
43 by some tests. If this is not provided, many tests will
45 (SCRATCH and TEST must be two DIFFERENT partitions)
47 - for btrfs only: some btrfs test cases will need 3 or more independent
48 SCRATCH disks which should be set using SCRATCH_DEV_POOL (for eg:
49 SCRATCH_DEV_POOL="/dev/sda /dev/sdb /dev/sdc") with which
50 SCRATCH_DEV should be unused by the tester, and for the legacy
51 support SCRATCH_DEV will be set to the first disk of the
52 SCRATCH_DEV_POOL by xfstests script.
54 - setup your environment
55 - setenv TEST_DEV "device containing TEST PARTITION"
56 - setenv TEST_DIR "mount point of TEST PARTITION"
58 - setenv SCRATCH_DEV "device containing SCRATCH PARTITION" OR
59 (btrfs only) setenv SCRATCH_DEV_POOL "to 3 or more SCRATCH disks for
60 testing btrfs raid concepts"
61 - setenv SCRATCH_MNT "mount point for SCRATCH PARTITION"
62 - setenv TAPE_DEV "tape device for testing xfsdump"
63 - setenv RMT_TAPE_DEV "remote tape device for testing xfsdump"
64 - setenv RMT_IRIXTAPE_DEV "remote IRIX tape device for testing xfsdump"
65 - setenv SCRATCH_LOGDEV "device for scratch-fs external log"
66 - setenv SCRATCH_RTDEV "device for scratch-fs realtime data"
67 - setenv TEST_LOGDEV "device for test-fs external log"
68 - setenv TEST_RTDEV "device for test-fs realtime data"
69 - if TEST_LOGDEV and/or TEST_RTDEV, these will always be used.
70 - if SCRATCH_LOGDEV and/or SCRATCH_RTDEV, the USE_EXTERNAL
71 environment variable set to "yes" will enable their use.
72 - setenv DIFF_LENGTH "number of diff lines to print from a failed test",
73 by default 10, set to 0 to print the full diff
74 - setenv FSTYP "the filesystem you want to test", the filesystem
75 type is devised from the TEST_DEV device, but you may want to
76 override it; if unset, the default is 'xfs'
77 - setenv FSSTRESS_AVOID and/or FSX_AVOID, which contain options
78 added to the end of fsstresss and fsx invocations, respectively,
79 in case you wish to exclude certain operational modes from these
82 - or add a case to the switch in common/config assigning
83 these variables based on the hostname of your test
85 - or add these variables to a file called local.config and keep that
86 file in your workarea.
88 - if testing xfsdump, make sure the tape devices have a
89 tape which can be overwritten.
91 - make sure $TEST_DEV is a mounted XFS partition
92 - make sure that $SCRATCH_DEV or $SCRATCH_DEV_POOL contains nothing useful
97 - By default the tests suite will run xfs tests:
98 - ./check '*/001' '*/002' '*/003'
100 - You can explicitly specify NFS/CIFS/UDF, otherwise the filesystem type will
101 be autodetected from $TEST_DEV:
102 ./check -nfs [test(s)]
103 - Groups of tests maybe ran by: ./check -g [group(s)]
104 See the 'group' file for details on groups
105 - for udf tests: ./check -udf [test(s)]
106 Running all the udf tests: ./check -udf -g udf
107 - for running nfs tests: ./check -nfs [test(s)]
108 - for running cifs/smb3 tests: ./check -cifs [test(s)]
109 - To randomize test order: ./check -r [test(s)]
112 The check script tests the return value of each script, and
113 compares the output against the expected output. If the output
114 is not as expected, a diff will be output and an .out.bad file
115 will be produced for the failing test.
117 Unexpected console messages, crashes and hangs may be considered
118 to be failures but are not necessarily detected by the QA system.
120 __________________________
121 ADDING TO THE FSQA SUITE
122 __________________________
125 Creating new tests scripts:
127 Use the "new" script.
129 Test script environment:
131 When developing a new test script keep the following things in
132 mind. All of the environment variables and shell procedures are
133 available to the script once the "common/rc" file has been
136 1. The tests are run from an arbitrary directory. If you want to
137 do operations on an XFS filesystem (good idea, eh?), then do
138 one of the following:
140 (a) Create directories and files at will in the directory
141 $TEST_DIR ... this is within an XFS filesystem and world
142 writeable. You should cleanup when your test is done,
143 e.g. use a _cleanup shell procedure in the trap ... see
144 001 for an example. If you need to know, the $TEST_DIR
145 directory is within the filesystem on the block device
148 (b) mkfs a new XFS filesystem on $SCRATCH_DEV, and mount this
149 on $SCRATCH_MNT. Call the the _require_scratch function
150 on startup if you require use of the scratch partition.
151 _require_scratch does some checks on $SCRATCH_DEV &
152 $SCRATCH_MNT and makes sure they're unmounted. You should
153 cleanup when your test is done, and in particular unmount
155 Tests can make use of $SCRATCH_LOGDEV and $SCRATCH_RTDEV
156 for testing external log and realtime volumes - however,
157 these tests need to simply "pass" (e.g. cat $seq.out; exit
158 - or default to an internal log) in the common case where
159 these variables are not set.
161 2. You can safely create temporary files that are not part of the
162 filesystem tests (e.g. to catch output, prepare lists of things
163 to do, etc.) in files named $tmp.<anything>. The standard test
164 script framework created by "new" will initialize $tmp and
167 3. By default, tests are run as the same uid as the person
168 executing the control script "check" that runs the test scripts.
170 If you need to be root, add a call to the shell procedure
171 _need_to_be_root ... this will do nothing or exit with an
172 error message depending on your current uid.
174 4. Some other useful shell procedures:
176 _get_fqdn - echo the host's fully qualified
179 _get_pids_by_name - one argument is a process name, and
180 return all of the matching pids on
183 _within_tolerance - fancy numerical "close enough is good
184 enough" filter for deterministic
185 output ... see comments in
186 common/filter for an explanation
188 _filter_date - turn ctime(3) format dates into the
189 string DATE for deterministic
192 _cat_passwd, - dump the content of the password
193 _cat_group or group file (both the local file
194 and the content of the NIS database
195 if it is likely to be present)
197 5. General recommendations, usage conventions, etc.:
198 - When the content of the password or group file is
199 required, get it using the _cat_passwd and _cat_group
200 functions, to ensure NIS information is included if NIS
202 - When calling getfacl in a test, pass the "-n" argument so
203 that numeric rather than symbolic identifiers are used in
208 Each test script has a numerical name, e.g. 007, and an associated
209 verified output, e.g. 007.out.
211 It is important that the verified output is deterministic, and
212 part of the job of the test script is to filter the output to
213 make this so. Examples of the sort of things that need filtering:
220 - variable directory contents
221 - imprecise numbers, especially sizes and times
225 The script "check" may be used to run one or more tests.
227 Test number $seq is deemed to "pass" when:
228 (a) no "core" file is created,
229 (b) the file $seq.notrun is not created,
230 (c) the exit status is 0, and
231 (d) the output matches the verified output.
233 In the "not run" case (b), the $seq.notrun file should contain a
234 short one-line summary of why the test was not run. The standard
235 output is not checked, so this can be used for a more verbose
236 explanation and to provide feedback when the QA test is run
240 To force a non-zero exit status use:
246 won't have the desired effect because of the way the exit trap
249 The recent pass/fail history is maintained in the file "check.log".
250 The elapsed time for the most recent pass for each test is kept