view tools/xm-test/README @ 7778:0b4596caf761

nloopbacks default is now 8. So vifnum of greater than 7 requires
an adjustment to nloopbacks. Warning comment updated.

Signed-off-by: Nivedita Singhvi (niv@us.ibm.com)
author kaf24@firebug.cl.cam.ac.uk
date Fri Nov 11 10:46:36 2005 +0100 (2005-11-11)
parents dfbdf1d264ef
children 9a8110cb09bd
line source
2 xm-test README
4 Copyright (C) International Business Machines Corp., 2005
5 Author(s): Dan Smith <danms@us.ibm.com>
6 Woody Marvel <marvel@us.ibm.com>
8 Overview
9 ========
11 This suite provides a framework for testing the Xen userspace tools.
12 The directory structure is:
14 ./xm-test
15 |
16 +-/lib: Python support libraries
17 |
18 +-/ramdisk: Staging area for building the test ramdisk
19 |
20 +-/tests
21 | |
22 | +-/create: Tests for the 'xm create' command
23 | +-/destroy: Tests for the 'xm destroy' command
24 | . . .
25 |
26 +-/utils: Utility scripts for ramdisk building
28 Reports are posted here:
30 http://xmtest.dague.org
33 Building
34 ========
36 Before the test suite can be used, the ramdisk must be built from
37 source. All source needed for this process is automatically
38 downloaded, extracted, and compiled. Due to the need to create
39 special files, this process must be done as root:
41 # ./configure
42 # make
44 NB: If you have the initrd.img from a previous version of xm-test, you
45 can copy it into the ramdisk directory to eliminate the need to
46 rebuild it. So far, nothing has changed in the ramdisk since xm-test
47 version 0.1.1. If you do this, there is no need to run 'make' again.
48 Simply copy the initrd.img file into ramdisk/ and then run the
49 runtest.sh script.
52 Running
53 =======
55 By running "./runtest.sh logfile" at the top level, all tests will be
56 run in alphabetic order. To run a specific test group, run "make
57 check" from inside that group directory. For example:
59 # cd tests/create
60 # make check
62 The runtest.sh script will create several files, including a .report
63 file, which is the cleaned up, email-friendly report of failures.
64 Additionally, the script will submit your results to the development
65 team for trend analysis. This helps us determine the level of success
66 people "out there" are having with different versions of Xen.
68 Note: you should generally run xm-test with a minimum of memory
69 allocated to Dom0. More memory available for allocation to DomUs
70 means a more rigorous test.
72 If you wish to run xm-test in an automated batch environment, you can
73 run the script with the "-b" flag, which will try to prevent it from
74 asking any questions interactively. You should run it manually at
75 least once to generate the "contact_info" file, which will be used in
76 subsequent runs.
78 BIG FAT WARNING: The framework assumes it is running on a dedicated
79 machine. As such, the library automatically destroys any running
80 DomUs on the sytem to provide each test with a "clean slate".
83 Extending
84 =========
86 Additional tests may be added in existing groups to test additional
87 cases for a given xm subcommand. Test programs should be named
88 according to the following scheme:
90 XY_group_name_{pos,neg}.py
92 Where:
93 XY is the next number in line
94 group is the name of the subcommand being tested
95 name is the short name of the test
96 {pos,neg} denotes whether this is a positive or negative test case
98 New subcommand groups should be added as directories named after the
99 subcommand itself. The "Makefile.am.template" should be copied into
100 the new group directory as "Makefile.am".
103 Developer Notes
104 ===============
106 Our library provides a DomU console abstraction for automated
107 execution of commands. Please note that this is relatively fragile,
108 and is intended for use only with the ramdisk built by the framework.
109 Because the console experiences some occasional corruption, this
110 method is not completely perfect at the moment, although the authors
111 use it with relatively few problems.
114 Known Issues
115 ============
117 - 08/22/2005 Dan Smith:
118 Current versions of Xen experience long delays on
119 boot, due to the blkif frontend waiting (and timing
120 out) for the backend. We now implicitly introduce
121 a pause in XmTestDomain.start() to remedy
122 this across all tests that need it.
124 - 09/22/2005 Dan Smith:
125 The save, restore, and migrate tests have been turned
126 back on, as we have implemented timeout functionality
127 in "traceCommand". Since migrate is known to hang, we
128 can now safely include it in the default test run, and
129 have it be killed if it never finishes. As before,
130 feedback about these tests would be specifically
131 appreciated.
133 Reporting Bugs
134 ==============
136 If you find a bug in the test framework, report it to:
138 Dan Smith <danms@us.ibm.com>
140 If you find a bug in a specific test case, contact the author of the
141 test case first.