PostgreSQL Regression Tests

The regression tests are a comprehensive set of tests for the SQL implementation in PostgreSQL. They test standard SQL operations as well as the extended capabilities of PostgreSQL.

Running the Tests

The regression tests can be run against an already installed and running server, or using a temporary installation within the build tree. Furthermore, there is a “parallel” and a “sequential” mode for running the tests. The sequential method runs each test script alone, while the parallel method starts up multiple server processes to run groups of tests in parallel. Parallel testing adds confidence that interprocess communication and locking are working correctly.

Running the Tests Against a Temporary Installation

To run the parallel regression tests after building but before installation, type:

make check

in the top-level directory. (Or you can change to src/test/regress and run the command there.) At the end you should see something like:

 
=======================
 All 193 tests passed.
=======================
 

or otherwise a note about which tests failed.

Because this test method runs a temporary server, it will not work if you did the build as the root user, since the server will not start as root. Recommended procedure is not to do the build as root, or else to perform testing after completing the installation.

If you have configured PostgreSQL to install into a location where an older PostgreSQL installation already exists, and you perform make check before installing the new version, you might find that the tests fail because the new programs try to use the already-installed shared libraries. (Typical symptoms are complaints about undefined symbols.) If you wish to run the tests before overwriting the old installation, you’ll need to build with configure --disable-rpath. It is not recommended that you use this option for the final installation, however.

The parallel regression test starts quite a few processes under your user ID. Presently, the maximum concurrency is twenty parallel test scripts, which means forty processes: there’s a server process and a psql process for each test script. So if your system enforces a per-user limit on the number of processes, make sure this limit is at least fifty or so, else you might get random-seeming failures in the parallel test. If you are not in a position to raise the limit, you can cut down the degree of parallelism by setting the MAX_CONNECTIONS parameter. For example:

make MAX_CONNECTIONS=10 check

runs no more than ten tests concurrently.

Running the Tests Against an Existing Installation

To run the tests after installation, initialize a data directory and start the server, then type:

make installcheck

or for a parallel test:

make installcheck-parallel

The tests will expect to contact the server at the local host and the default port number, unless directed otherwise by PGHOST and PGPORT environment variables. The tests will be run in a database named regression; any existing database by this name will be dropped.

The tests will also transiently create some cluster-wide objects, such as roles, tablespaces, and subscriptions. These objects will have names beginning with regress_. Beware of using installcheck mode with an installation that has any actual global objects named that way.

Additional Test Suites

The make check and make installcheck commands run only the “core” regression tests, which test built-in functionality of the PostgreSQL server. The source distribution contains many additional test suites, most of them having to do with add-on functionality such as optional procedural languages.

To run all test suites applicable to the modules that have been selected to be built, including the core tests, type one of these commands at the top of the build tree:

make check-world
make installcheck-world

These commands run the tests using temporary servers or an already-installed server, respectively, just as previously explained for make check and make installcheck. Other considerations are the same as previously explained for each method. Note that make check-world builds a separate instance (temporary data directory) for each tested module, so it requires more time and disk space than make installcheck-world.

On a modern machine with multiple CPU cores and no tight operating-system limits, you can make things go substantially faster with parallelism. The recipe that most PostgreSQL developers actually use for running all tests is something like

make check-world -j8 >/dev/null

with a -j limit near to or a bit more than the number of available cores. Discarding stdout eliminates chatter that’s not interesting when you just want to verify success. (In case of failure, the stderr messages are usually enough to determine where to look closer.)

Alternatively, you can run individual test suites by typing make check or make installcheck in the appropriate subdirectory of the build tree. Keep in mind that make installcheck assumes you’ve installed the relevant module(s), not only the core server.

The additional tests that can be invoked this way include:

  • Regression tests for optional procedural languages. These are located under src/pl.
  • Regression tests for contrib modules, located under contrib. Not all contrib modules have tests.
  • Regression tests for the ECPG interface library, located in src/interfaces/ecpg/test.
  • Tests for core-supported authentication methods, located in src/test/authentication. (See below for additional authentication-related tests.)
  • Tests stressing behavior of concurrent sessions, located in src/test/isolation.
  • Tests for crash recovery and physical replication, located in src/test/recovery.
  • Tests for logical replication, located in src/test/subscription.
  • Tests of client programs, located under src/bin.

When using installcheck mode, these tests will create and destroy test databases whose names include regression, for example pl_regression or contrib_regression. Beware of using installcheck mode with an installation that has any non-test databases named that way.

Some of these auxiliary test suites use the TAP infrastructure. The TAP-based tests are run only when PostgreSQL was configured with the option --enable-tap-tests. This is recommended for development, but can be omitted if there is no suitable Perl installation.

Some test suites are not run by default, either because they are not secure to run on a multiuser system or because they require special software. You can decide which test suites to run additionally by setting the make or environment variable PG_TEST_EXTRA to a whitespace-separated list, for example:

make check-world PG_TEST_EXTRA='kerberos ldap ssl'

The following values are currently supported:

kerberos

Runs the test suite under src/test/kerberos. This requires an MIT Kerberos installation and opens TCP/IP listen sockets.

ldap

Runs the test suite under src/test/ldap. This requires an OpenLDAP installation and opens TCP/IP listen sockets.

ssl

Runs the test suite under src/test/ssl. This opens TCP/IP listen sockets.

Tests for features that are not supported by the current build configuration are not run even if they are mentioned in PG_TEST_EXTRA.

In addition, there are tests in src/test/modules which will be run by make check-world but not by make installcheck-world. This is because they install non-production extensions or have other side-effects that are considered undesirable for a production installation. You can use make install and make installcheck in one of those subdirectories if you wish, but it’s not recommended to do so with a non-test server.

Locale and Encoding

By default, tests using a temporary installation use the locale defined in the current environment and the corresponding database encoding as determined by initdb. It can be useful to test different locales by setting the appropriate environment variables, for example:

make check LANG=C
make check LC_COLLATE=en_US.utf8 LC_CTYPE=fr_CA.utf8

For implementation reasons, setting LC_ALL does not work for this purpose; all the other locale-related environment variables do work.

When testing against an existing installation, the locale is determined by the existing database cluster and cannot be set separately for the test run.

You can also choose the database encoding explicitly by setting the variable ENCODING, for example:

make check LANG=C ENCODING=EUC_JP

Setting the database encoding this way typically only makes sense if the locale is C; otherwise the encoding is chosen automatically from the locale, and specifying an encoding that does not match the locale will result in an error.

The database encoding can be set for tests against either a temporary or an existing installation, though in the latter case it must be compatible with the installation’s locale.

Extra Tests

The core regression test suite contains a few test files that are not run by default, because they might be platform-dependent or take a very long time to run. You can run these or other extra test files by setting the variable EXTRA_TESTS. For example, to run the numeric_big test:

make check EXTRA_TESTS=numeric_big

Testing Hot Standby

The source distribution also contains regression tests for the static behavior of Hot Standby. These tests require a running primary server and a running standby server that is accepting new WAL changes from the primary (using either file-based log shipping or streaming replication). Those servers are not automatically created for you, nor is replication setup documented here. Please check the various sections of the documentation devoted to the required commands and related issues.

To run the Hot Standby tests, first create a database called regression on the primary:

psql -h primary -c "CREATE DATABASE regression"

Next, run the preparatory script src/test/regress/sql/hs_primary_setup.sql on the primary in the regression database, for example:

psql -h primary -f src/test/regress/sql/hs_primary_setup.sql regression

Allow these changes to propagate to the standby.

Now arrange for the default database connection to be to the standby server under test (for example, by setting the PGHOST and PGPORT environment variables). Finally, run make standbycheck in the regression directory:

cd src/test/regress
make standbycheck

Some extreme behaviors can also be generated on the primary using the script src/test/regress/sql/hs_primary_extremes.sql to allow the behavior of the standby to be tested.

Test Evaluation

Some properly installed and fully functional PostgreSQL installations can “fail” some of these regression tests due to platform-specific artifacts such as varying floating-point representation and message wording. The tests are currently evaluated using a simple diff comparison against the outputs generated on a reference system, so the results are sensitive to small system differences. When a test is reported as “failed”, always examine the differences between expected and actual results; you might find that the differences are not significant. Nonetheless, we still strive to maintain accurate reference files across all supported platforms, so it can be expected that all tests pass.

The actual outputs of the regression tests are in files in the src/test/regress/results directory. The test script uses diff to compare each output file against the reference outputs stored in the src/test/regress/expected directory. Any differences are saved for your inspection in src/test/regress/regression.diffs. (When running a test suite other than the core tests, these files of course appear in the relevant subdirectory, not src/test/regress.)

If you don’t like the diff options that are used by default, set the environment variable PG_REGRESS_DIFF_OPTS, for instance PG_REGRESS_DIFF_OPTS='-c'. (Or you can run diff yourself, if you prefer.)

If for some reason a particular platform generates a “failure” for a given test, but inspection of the output convinces you that the result is valid, you can add a new comparison file to silence the failure report in future test runs.

Error Message Differences

Some of the regression tests involve intentional invalid input values. Error messages can come from either the PostgreSQL code or from the host platform system routines. In the latter case, the messages can vary between platforms, but should reflect similar information. These differences in messages will result in a “failed” regression test that can be validated by inspection.

Locale Differences

If you run the tests against a server that was initialized with a collation-order locale other than C, then there might be differences due to sort order and subsequent failures. The regression test suite is set up to handle this problem by providing alternate result files that together are known to handle a large number of locales.

To run the tests in a different locale when using the temporary-installation method, pass the appropriate locale-related environment variables on the make command line, for example:

make check LANG=de_DE.utf8

(The regression test driver unsets LC_ALL, so it does not work to choose the locale using that variable.) To use no locale, either unset all locale-related environment variables (or set them to C) or use the following special invocation:

make check NO_LOCALE=1

When running the tests against an existing installation, the locale setup is determined by the existing installation. To change it, initialize the database cluster with a different locale by passing the appropriate options to initdb.

In general, it is advisable to try to run the regression tests in the locale setup that is wanted for production use, as this will exercise the locale- and encoding-related code portions that will actually be used in production. Depending on the operating system environment, you might get failures, but then you will at least know what locale-specific behaviors to expect when running real applications.

Date and Time Differences

Most of the date and time results are dependent on the time zone environment. The reference files are generated for time zone PST8PDT (Berkeley, California), and there will be apparent failures if the tests are not run with that time zone setting. The regression test driver sets environment variable PGTZ to PST8PDT, which normally ensures proper results.

Floating-Point Differences

Some of the tests involve computing 64-bit floating-point numbers (double precision) from table columns. Differences in results involving mathematical functions of double precision columns have been observed. The float8 and geometry tests are particularly prone to small differences across platforms, or even with different compiler optimization settings. Human eyeball comparison is needed to determine the real significance of these differences which are usually 10 places to the right of the decimal point.

Some systems display minus zero as -0, while others just show 0.

Some systems signal errors from pow() and exp() differently from the mechanism expected by the current PostgreSQL code.

Row Ordering Differences

You might see differences in which the same rows are output in a different order than what appears in the expected file. In most cases this is not, strictly speaking, a bug. Most of the regression test scripts are not so pedantic as to use an ORDER BY for every single SELECT, and so their result row orderings are not well-defined according to the SQL specification. In practice, since we are looking at the same queries being executed on the same data by the same software, we usually get the same result ordering on all platforms, so the lack of ORDER BY is not a problem. Some queries do exhibit cross-platform ordering differences, however. When testing against an already-installed server, ordering differences can also be caused by non-C locale settings or non-default parameter settings, such as custom values of work_mem or the planner cost parameters.

Therefore, if you see an ordering difference, it’s not something to worry about, unless the query does have an ORDER BY that your result is violating. However, please report it anyway, so that we can add an ORDER BY to that particular query to eliminate the bogus “failure” in future releases.

You might wonder why we don’t order all the regression test queries explicitly to get rid of this issue once and for all. The reason is that that would make the regression tests less useful, not more, since they’d tend to exercise query plan types that produce ordered results to the exclusion of those that don’t.

Insufficient Stack Depth

If the errors test results in a server crash at the select infinite_recurse() command, it means that the platform’s limit on process stack size is smaller than the max_stack_depth parameter indicates. This can be fixed by running the server under a higher stack size limit (4MB is recommended with the default value of max_stack_depth). If you are unable to do that, an alternative is to reduce the value of max_stack_depth.

On platforms supporting getrlimit(), the server should automatically choose a safe value of max_stack_depth; so unless you’ve manually overridden this setting, a failure of this kind is a reportable bug.

The “random” Test

The random test script is intended to produce random results. In very rare cases, this causes that regression test to fail. Typing:

diff results/random.out expected/random.out

should produce only one or a few lines of differences. You need not worry unless the random test fails repeatedly.

Configuration Parameters

When running the tests against an existing installation, some non-default parameter settings could cause the tests to fail. For example, changing parameters such as enable_seqscan or enable_indexscan could cause plan changes that would affect the results of tests that use EXPLAIN.

Variant Comparison Files

Since some of the tests inherently produce environment-dependent results, we have provided ways to specify alternate “expected” result files. Each regression test can have several comparison files showing possible results on different platforms. There are two independent mechanisms for determining which comparison file is used for each test.

The first mechanism allows comparison files to be selected for specific platforms. There is a mapping file, src/test/regress/resultmap, that defines which comparison file to use for each platform. To eliminate bogus test “failures” for a particular platform, you first choose or make a variant result file, and then add a line to the resultmap file.

Each line in the mapping file is of the form

testname:output:platformpattern=comparisonfilename

The test name is just the name of the particular regression test module. The output value indicates which output file to check. For the standard regression tests, this is always out. The value corresponds to the file extension of the output file. The platform pattern is a pattern in the style of the Unix tool expr (that is, a regular expression with an implicit ^ anchor at the start). It is matched against the platform name as printed by config.guess. The comparison file name is the base name of the substitute result comparison file.

For example: some systems lack a working strtof function, for which our workaround causes rounding errors in the float4 regression test. Therefore, we provide a variant comparison file, float4-misrounded-input.out, which includes the results to be expected on these systems. To silence the bogus “failure” message on HP-UX 10 platforms, resultmap includes:

float4:out:hppa.*-hp-hpux10.*=float4-misrounded-input.out

which will trigger on any machine where the output of config.guess matches hppa.*-hp-hpux10.*. Other lines in resultmap select the variant comparison file for other platforms where it’s appropriate.

The second selection mechanism for variant comparison files is much more automatic: it simply uses the “best match” among several supplied comparison files. The regression test driver script considers both the standard comparison file for a test, testname.out, and variant files named testname_digit.out (where the digit is any single digit 09). If any such file is an exact match, the test is considered to pass; otherwise, the one that generates the shortest diff is used to create the failure report. (If resultmap includes an entry for the particular test, then the base testname is the substitute name given in resultmap.)

For example, for the char test, the comparison file char.out contains results that are expected in the C and POSIX locales, while the file char_1.out contains results sorted as they appear in many other locales.

The best-match mechanism was devised to cope with locale-dependent results, but it can be used in any situation where the test results cannot be predicted easily from the platform name alone. A limitation of this mechanism is that the test driver cannot tell which variant is actually “correct” for the current environment; it will just pick the variant that seems to work best. Therefore it is safest to use this mechanism only for variant results that you are willing to consider equally valid in all contexts.

TAP Tests

Various tests, particularly the client program tests under src/bin, use the Perl TAP tools and are run using the Perl testing program prove. You can pass command-line options to prove by setting the make variable PROVE_FLAGS, for example:

make -C src/bin check PROVE_FLAGS='--timer'

See the manual page of prove for more information.

The make variable PROVE_TESTS can be used to define a whitespace-separated list of paths relative to the Makefile invoking prove to run the specified subset of tests instead of the default t/*.pl. For example:

make check PROVE_TESTS='t/001_test1.pl t/003_test3.pl'

The TAP tests require the Perl module IPC::Run. This module is available from CPAN or an operating system package.

Generically speaking, the TAP tests will test the executables in a previously-installed installation tree if you say make installcheck, or will build a new local installation tree from current sources if you say make check. In either case they will initialize a local instance (data directory) and transiently run a server in it. Some of these tests run more than one server. Thus, these tests can be fairly resource-intensive.

It’s important to realize that the TAP tests will start test server(s) even when you say make installcheck; this is unlike the traditional non-TAP testing infrastructure, which expects to use an already-running test server in that case. Some PostgreSQL subdirectories contain both traditional-style and TAP-style tests, meaning that make installcheck will produce a mix of results from temporary servers and the already-running test server.

Test Coverage Examination

The PostgreSQL source code can be compiled with coverage testing instrumentation, so that it becomes possible to examine which parts of the code are covered by the regression tests or any other test suite that is run with the code. This is currently supported when compiling with GCC and requires the gcov and lcov programs.

A typical workflow would look like this:

./configure --enable-coverage ... OTHER OPTIONS ...
make
make check # or other test suite
make coverage-html

Then point your HTML browser to coverage/index.html. The make commands also work in subdirectories.

If you don’t have lcov or prefer text output over an HTML report, you can also run

make coverage

instead of make coverage-html, which will produce .gcov output files for each source file relevant to the test. (make coverage and make coverage-html will overwrite each other’s files, so mixing them might be confusing.)

To reset the execution counts between test runs, run:

make coverage-clean