Elena Stepanova added a comment - 2017-02-13 00:55 In theory, it would be nice to be able to run all MTR tests with an arbitrary set of non-default options and get clean results (there have been a discussion on a mailing list about it, too). In reality, generally speaking, we can't do it. The "clean results" part is important here. It is always possible to do such ad-hoc runs experimentally, to see that there are no crashes and that differences produced by tests look reasonable; but of course this distinction would have to be done by a human. If we want to automate these runs as suggested, they should actually produce the exact expected results, which not only means that existing tests should be modified, and they will inevitably lose some precision (often it's redundant precision, but it might not always be so), but, importantly, it also means that new tests will have to be developered with the consideration of these non-default runs, which can never happen (especially if they are not run per-push, and we can't afford running them per-push), thus new tests will be failing regularly on these runs, which makes maintenance more difficult. We have hundreds of server options, and if we want to adapt tests to all of them, tests will become meaningless. It's the limitation of the result-based nature of MTR that it checks not only the exact thing that is being tested, but lots of unrelated artefacts. If there is a strong demand and a good excuse for regular running test suites with non-default values of some particular option, it can probably be done. But normally a more reasonable approach is to choose tests which are really important for testing this particular option, and make them run with the desired values of this option (e.g. via 'combinations' facility). Now, it is impossible to run combinations with innodb_page_size , because its non-default values are incompatible with normal MTR's bootstrap. MTR's bootstrap has already been made aware of --innodb* command-line options, which is why it is now possible to run tests with different innodb_page_size values provided via the command line (it hadn't always been so). But MTR still ignores options from opt files and combinations files, even although they also end up on mysqld command line (as opposed to options in cnf files which go to the config). So, I suggest to start by fixing this. Currently MTR bootstraps the server once, it creates and stores a clean datadir. Then, when the server needs to be restarted between tests, MTR copies the previously stored datadir and starts the server on the copy. What we need to do is in certain cases make it run bootstrap again, directly on the target datadir, without touching the default clean one, and use the command-line options for this bootstrap. We don't want to do it every time, it would be way too expensive, and only a tiny portion of tests actually needs it. So, the suggestion is to parse the command-line options to check for those which require re-bootstrap, and do it only if any of them are present. It is a bit dirty as it requires hardcoding, but in fact there is nothing particularly new there, MTR does all kinds of dirty tricks like that. Here is a proposed change in MTR (it's much smaller than the text above): https://github.com/MariaDB/server/commit/f52be90cc91d9effe655f73352289fe85a029f67 And here is a test case with innodb_page_size combinations: https://github.com/MariaDB/server/commit/51c8b39db65b1b97f5281467eba5652a50fce08d It's a useless test only made as an example. It intentionally produces different results with different page sizes, to demonstrate how combinations and rdiff files work together.

Elena Stepanova added a comment - 2017-02-13 00:57 marko , Please check the suggestion above and the MTR change. If you are okay with doing it this way, let's stick to it for now. If you still want to run all suites with the option, please re-assign it back to me (and let me know whether you want the change in MTR anyway, or not).

Marko Mäkelä added a comment - 2017-02-24 09:48 elenst , I like your approach. I see that you even worked around the minimum buffer pool size requirement. We should lift that restriction in 10.2 and merge the fix for Bug#22179133 INNODB: TOO SMALL BUFFER POOL FOR INNODB_PAGE_SIZE=64K .

Elena Stepanova added a comment - 2017-02-24 10:45 - edited Right. I don't remember why I didn't do it in the tree, but naturally at the end it should be innodb_page_size.inc and innodb_page_size.combinations, and any test which wants to be run with different page sizes should just include it (and possibly add rdiffs, if it has some output which is different for different page sizes, but it's not necessarily so). The combinations file will take care of additional options. Like this: diff --git a/mysql-test/include/innodb_page_size.combinations b/mysql-test/include/innodb_page_size.combinations new file mode 100644 index 0000000..a76431e --- /dev/null +++ b/mysql-test/include/innodb_page_size.combinations @@ -0,0 +1,16 @@ +[32k] +innodb-page-size=32K +innodb-buffer-pool-size=32M + +[16k] +innodb-page-size=16K + +[64k] +innodb-page-size=64K +innodb-buffer-pool-size=64M + +[8k] +innodb-page-size=8K + +[4k] +innodb-page-size=4K diff --git a/mysql-test/include/innodb_page_size.inc b/mysql-test/include/innodb_page_size.inc new file mode 100644 index 0000000..754b640 --- /dev/null +++ b/mysql-test/include/innodb_page_size.inc @@ -0,0 +1,4 @@ +# The goal of including this file is to enable innodb_page_size combinations +# (see include/innodb_page_size.combinations) + +--source include/have_innodb.inc diff --git a/mysql-test/suite/innodb/r/innodb_page_size,32k.rdiff b/mysql-test/suite/innodb/r/innodb_page_size,32k.rdiff new file mode 100644 index 0000000..eb3d888 --- /dev/null +++ b/mysql-test/suite/innodb/r/innodb_page_size,32k.rdiff @@ -0,0 +1,7 @@ +--- /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.result 2017-02-10 17:19:35.998098072 +0200 ++++ /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.reject 2017-02-10 17:27:02.770095902 +0200 +@@ -1,3 +1,3 @@ + select @@innodb_page_size; + @@innodb_page_size +-16384 ++32768 diff --git a/mysql-test/suite/innodb/r/innodb_page_size,4k.rdiff b/mysql-test/suite/innodb/r/innodb_page_size,4k.rdiff new file mode 100644 index 0000000..7db0ff7 --- /dev/null +++ b/mysql-test/suite/innodb/r/innodb_page_size,4k.rdiff @@ -0,0 +1,7 @@ +--- /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.result 2017-02-10 17:19:35.998098072 +0200 ++++ /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.reject 2017-02-10 17:27:02.770095902 +0200 +@@ -1,3 +1,3 @@ + select @@innodb_page_size; + @@innodb_page_size +-16384 ++4096 diff --git a/mysql-test/suite/innodb/r/innodb_page_size,64k.rdiff b/mysql-test/suite/innodb/r/innodb_page_size,64k.rdiff new file mode 100644 index 0000000..5a90fcc --- /dev/null +++ b/mysql-test/suite/innodb/r/innodb_page_size,64k.rdiff @@ -0,0 +1,7 @@ +--- /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.result 2017-02-10 17:19:35.998098072 +0200 ++++ /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.reject 2017-02-10 17:27:02.770095902 +0200 +@@ -1,3 +1,3 @@ + select @@innodb_page_size; + @@innodb_page_size +-16384 ++65536 diff --git a/mysql-test/suite/innodb/r/innodb_page_size,8k.rdiff b/mysql-test/suite/innodb/r/innodb_page_size,8k.rdiff new file mode 100644 index 0000000..6c59897 --- /dev/null +++ b/mysql-test/suite/innodb/r/innodb_page_size,8k.rdiff @@ -0,0 +1,7 @@ +--- /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.result 2017-02-10 17:19:35.998098072 +0200 ++++ /data/src/bb-10.2-elenst/mysql-test/suite/innodb/r/innodb_page_size.reject 2017-02-10 17:27:02.770095902 +0200 +@@ -1,3 +1,3 @@ + select @@innodb_page_size; + @@innodb_page_size +-16384 ++8192 diff --git a/mysql-test/suite/innodb/r/innodb_page_size.result b/mysql-test/suite/innodb/r/innodb_page_size.result new file mode 100644 index 0000000..349ec4b --- /dev/null +++ b/mysql-test/suite/innodb/r/innodb_page_size.result @@ -0,0 +1,3 @@ +select @@innodb_page_size; +@@innodb_page_size +16384 diff --git a/mysql-test/suite/innodb/t/innodb_page_size.test b/mysql-test/suite/innodb/t/innodb_page_size.test new file mode 100644 index 0000000..c9b16fb --- /dev/null +++ b/mysql-test/suite/innodb/t/innodb_page_size.test @@ -0,0 +1,6 @@ +--source include/innodb_page_size.inc + +# This test intentionally produces different output +# depending on innodb_page_size, to demonstrate how rdiffs work + +select @@innodb_page_size; diff --git a/mysql-test/suite/innodb_fts/t/create.test b/mysql-test/suite/innodb_fts/t/create.test index fe0273a..49d0ecd 100644 --- a/mysql-test/suite/innodb_fts/t/create.test +++ b/mysql-test/suite/innodb_fts/t/create.test @@ -1,4 +1,5 @@ --source include/have_innodb.inc +--source include/innodb_page_size.inc SET NAMES utf8mb4; --echo # Here the new innodb_page_size.test demonstrates how it works when the output depends on innodb_page_size (rdiffs are needed), and the old innodb_fts.create test demonstrates how it works when the output is the same (no changes on the result side). $ perl ./mtr innodb.innodb_page_size innodb_fts.create   worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 innodb_fts.create '32k,innodb' [ pass ] 9265 innodb.innodb_page_size '32k,innodb' [ pass ] 2 innodb_fts.create '64k,innodb' [ pass ] 9875 innodb.innodb_page_size '64k,innodb' [ pass ] 3 innodb_fts.create '16k,innodb' [ pass ] 9151 innodb.innodb_page_size '16k,innodb' [ pass ] 3 innodb_fts.create '4k,innodb' [ pass ] 9460 innodb.innodb_page_size '4k,innodb' [ pass ] 2 innodb_fts.create '8k,innodb' [ pass ] 9287 innodb.innodb_page_size '8k,innodb' [ pass ] 2 -------------------------------------------------------------------------- The servers were restarted 9 times Spent 47.050 of 178 seconds executing testcases Please remember that even 5 combinations isn't cheap, especially with rebootstraps, and if a test uses orthogonal combinations (e.g. binlog_format ones), the number of test runs will multiply. Use it wisely.

Marko Mäkelä added a comment - 2017-05-29 12:33 Looks OK to me.

Elena Stepanova added a comment - 2017-06-02 11:13 - edited A suitable part of the change is pushed into 10.0: https://github.com/MariaDB/server/commit/ab62b7538f2eab43943da56c3a6d74cd85dfa877 Please note that in this scope of this task only the facility to run tests with innodb_page_size configurations is provided; the tests themselves have not been modified. To be run with different innodb_page_size values, a test should call --source include/innodb_page_size.inc Additional changes are required after merging to 10.1, since 10.0 part only adds combinations for page size 4/8/16K, while 10.1 also needs 32/64K. The task will remain open until the 10.1 part is done.

Elena Stepanova added a comment - 2017-06-15 23:16 Marko has added 32/64k combinations in 10.1 after the merge, so it can be closed now.