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")
15 - create 123456-fsgqa test user ("sudo useradd 123456-fsgqa")
18 - cd into the xfstests directory
19 - set the ROOT and TOOLROOT env variables for IRIX appropriately
22 ______________________
24 ______________________
26 Preparing system for tests (IRIX and Linux):
28 - compile XFS into your kernel or load XFS modules
29 - install user tools including mkfs.xfs, xfs_db & xfs_bmap
30 - If you wish to run the udf components of the suite install
31 mkfs_udf and udf_db for IRIX and mkudffs for Linux. Also download and
32 build the Philips UDF Verification Software from
33 http://www.extra.research.philips.com/udf/, then copy the udf_test
34 binary to xfstests/src/. If you wish to disable UDF verification test
35 set the environment variable DISABLE_UDF_TEST to 1.
38 - create one or two partitions to use for testing
40 - format as XFS, mount & optionally populate with
42 - one SCRATCH partition (optional)
43 - leave empty and expect this partition to be clobbered
44 by some tests. If this is not provided, many tests will
46 (SCRATCH and TEST must be two DIFFERENT partitions)
48 - for btrfs only: some btrfs test cases will need 3 or more independent
49 SCRATCH disks which should be set using SCRATCH_DEV_POOL (for eg:
50 SCRATCH_DEV_POOL="/dev/sda /dev/sdb /dev/sdc") with which
51 SCRATCH_DEV should be unused by the tester, and for the legacy
52 support SCRATCH_DEV will be set to the first disk of the
53 SCRATCH_DEV_POOL by xfstests script.
55 - setup your environment
56 - setenv TEST_DEV "device containing TEST PARTITION"
57 - setenv TEST_DIR "mount point of TEST PARTITION"
59 - setenv SCRATCH_DEV "device containing SCRATCH PARTITION" OR
60 (btrfs only) setenv SCRATCH_DEV_POOL "to 3 or more SCRATCH disks for
61 testing btrfs raid concepts"
62 - setenv SCRATCH_MNT "mount point for SCRATCH PARTITION"
63 - setenv TAPE_DEV "tape device for testing xfsdump"
64 - setenv RMT_TAPE_DEV "remote tape device for testing xfsdump"
65 - setenv RMT_IRIXTAPE_DEV "remote IRIX tape device for testing xfsdump"
66 - setenv SCRATCH_LOGDEV "device for scratch-fs external log"
67 - setenv SCRATCH_RTDEV "device for scratch-fs realtime data"
68 - setenv TEST_LOGDEV "device for test-fs external log"
69 - setenv TEST_RTDEV "device for test-fs realtime data"
70 - if TEST_LOGDEV and/or TEST_RTDEV, these will always be used.
71 - if SCRATCH_LOGDEV and/or SCRATCH_RTDEV, the USE_EXTERNAL
72 environment variable set to "yes" will enable their use.
73 - setenv DIFF_LENGTH "number of diff lines to print from a failed test",
74 by default 10, set to 0 to print the full diff
75 - setenv FSTYP "the filesystem you want to test", the filesystem
76 type is devised from the TEST_DEV device, but you may want to
77 override it; if unset, the default is 'xfs'
78 - setenv FSSTRESS_AVOID and/or FSX_AVOID, which contain options
79 added to the end of fsstresss and fsx invocations, respectively,
80 in case you wish to exclude certain operational modes from these
83 - or add a case to the switch in common/config assigning
84 these variables based on the hostname of your test
86 - or add these variables to a file called local.config and keep that
87 file in your workarea.
89 - if testing xfsdump, make sure the tape devices have a
90 tape which can be overwritten.
92 - make sure $TEST_DEV is a mounted XFS partition
93 - make sure that $SCRATCH_DEV or $SCRATCH_DEV_POOL contains nothing useful
98 - By default the tests suite will run xfs tests:
99 - ./check '*/001' '*/002' '*/003'
101 - You can explicitly specify NFS/CIFS/UDF, otherwise the filesystem type will
102 be autodetected from $TEST_DEV:
103 ./check -nfs [test(s)]
104 - Groups of tests maybe ran by: ./check -g [group(s)]
105 See the 'group' file for details on groups
106 - for udf tests: ./check -udf [test(s)]
107 Running all the udf tests: ./check -udf -g udf
108 - for running nfs tests: ./check -nfs [test(s)]
109 - for running cifs/smb3 tests: ./check -cifs [test(s)]
110 - To randomize test order: ./check -r [test(s)]
113 The check script tests the return value of each script, and
114 compares the output against the expected output. If the output
115 is not as expected, a diff will be output and an .out.bad file
116 will be produced for the failing test.
118 Unexpected console messages, crashes and hangs may be considered
119 to be failures but are not necessarily detected by the QA system.
121 __________________________
122 ADDING TO THE FSQA SUITE
123 __________________________
126 Creating new tests scripts:
128 Use the "new" script.
130 Test script environment:
132 When developing a new test script keep the following things in
133 mind. All of the environment variables and shell procedures are
134 available to the script once the "common/rc" file has been
137 1. The tests are run from an arbitrary directory. If you want to
138 do operations on an XFS filesystem (good idea, eh?), then do
139 one of the following:
141 (a) Create directories and files at will in the directory
142 $TEST_DIR ... this is within an XFS filesystem and world
143 writeable. You should cleanup when your test is done,
144 e.g. use a _cleanup shell procedure in the trap ... see
145 001 for an example. If you need to know, the $TEST_DIR
146 directory is within the filesystem on the block device
149 (b) mkfs a new XFS filesystem on $SCRATCH_DEV, and mount this
150 on $SCRATCH_MNT. Call the the _require_scratch function
151 on startup if you require use of the scratch partition.
152 _require_scratch does some checks on $SCRATCH_DEV &
153 $SCRATCH_MNT and makes sure they're unmounted. You should
154 cleanup when your test is done, and in particular unmount
156 Tests can make use of $SCRATCH_LOGDEV and $SCRATCH_RTDEV
157 for testing external log and realtime volumes - however,
158 these tests need to simply "pass" (e.g. cat $seq.out; exit
159 - or default to an internal log) in the common case where
160 these variables are not set.
162 2. You can safely create temporary files that are not part of the
163 filesystem tests (e.g. to catch output, prepare lists of things
164 to do, etc.) in files named $tmp.<anything>. The standard test
165 script framework created by "new" will initialize $tmp and
168 3. By default, tests are run as the same uid as the person
169 executing the control script "check" that runs the test scripts.
171 If you need to be root, add a call to the shell procedure
172 _need_to_be_root ... this will do nothing or exit with an
173 error message depending on your current uid.
175 4. Some other useful shell procedures:
177 _get_fqdn - echo the host's fully qualified
180 _get_pids_by_name - one argument is a process name, and
181 return all of the matching pids on
184 _within_tolerance - fancy numerical "close enough is good
185 enough" filter for deterministic
186 output ... see comments in
187 common/filter for an explanation
189 _filter_date - turn ctime(3) format dates into the
190 string DATE for deterministic
193 _cat_passwd, - dump the content of the password
194 _cat_group or group file (both the local file
195 and the content of the NIS database
196 if it is likely to be present)
198 5. General recommendations, usage conventions, etc.:
199 - When the content of the password or group file is
200 required, get it using the _cat_passwd and _cat_group
201 functions, to ensure NIS information is included if NIS
203 - When calling getfacl in a test, pass the "-n" argument so
204 that numeric rather than symbolic identifiers are used in
206 - When creating a new test, it is possible to enter a custom name
207 for the file. Filenames are in form NNN-custom-name, where NNN
208 is automatically added by the ./new script as an unique ID,
209 and "custom-name" is the optional string entered into a prompt
210 in the ./new script. It can contain only alphanumeric characters
211 and dash. Note the "NNN-" part is added automatically.
215 Each test script has a name, e.g. 007, and an associated
216 verified output, e.g. 007.out.
218 It is important that the verified output is deterministic, and
219 part of the job of the test script is to filter the output to
220 make this so. Examples of the sort of things that need filtering:
227 - variable directory contents
228 - imprecise numbers, especially sizes and times
232 The script "check" may be used to run one or more tests.
234 Test number $seq is deemed to "pass" when:
235 (a) no "core" file is created,
236 (b) the file $seq.notrun is not created,
237 (c) the exit status is 0, and
238 (d) the output matches the verified output.
240 In the "not run" case (b), the $seq.notrun file should contain a
241 short one-line summary of why the test was not run. The standard
242 output is not checked, so this can be used for a more verbose
243 explanation and to provide feedback when the QA test is run
247 To force a non-zero exit status use:
253 won't have the desired effect because of the way the exit trap
256 The recent pass/fail history is maintained in the file "check.log".
257 The elapsed time for the most recent pass for each test is kept