percona/percona-toolkit
1
0
Fork 0
mirror of https://github.com/percona/percona-toolkit.git synced 2025-09-15 16:00:23 +00:00
Code Issues Packages Projects Releases Wiki Activity
1,118 Commits 144 Branches 71 Tags
79c716d3586b63f99ffda71b95ccd0557f277ea5
Commit Graph

352 Commits

Author SHA1 Message Date
Daniel Nichter
79c716d358 Remove pt-table-sync/issue_79.t (it was broken but coincidentally working) and write better tests in filters.t. Add subs to Sandbox and PerconaTest to clear and parse tables used from the general logs. Set log=genlog in the sandbox configs. 2012-06-07 12:38:45 -04:00
Brian Fraser fraserb@gmail.com
03a9461daf Removed PerconaTest::{wait_until_no_lag|wait_until_slaves_running}, replaced them with Sandbox::wait_for_slaves 2012-06-07 12:00:40 -03:00
Baron Schwartz
af38351e82 wait for slaves after load_file 2012-06-06 16:32:14 -04:00
Daniel Nichter
f4b3bdb88b Use just RAND() instead of RAND() + NOW() to avoid be coerced to an int and therefore not sub-second safe. 2012-06-06 13:52:17 -07:00
Daniel Nichter
a178bb32f8 Use md5 "pings" in percona_test.sentinel for sub-second tests that call wipe_clean() twice. Update t/lib/samples/SchemaIterator/all-dbs-tbls-5.0.txt for percona_test.sentinel. Restart slaves in pt-table-checksum/replication_filters.t. Modernize and make pt-table-sync/replicate_do_db.t stable. 2012-06-06 13:31:24 -07:00
Daniel Nichter
14b615ebd9 5s default wait_until() time. Wait 5m instead of 16m for wipe_clean(). Don't connect to slave2 in wipe_clean() any longer. 2012-06-06 11:43:56 -07:00
Baron Schwartz
9865ad451d add sanity checks and slave-lag synchronization to prevent replication breakage, false sandbox breakage, etc. 2012-06-06 04:47:45 -04:00
Brian Fraser fraserb@gmail.com
c2bb24befc Merged lp:~percona-toolkit-dev/percona-toolkit/replace-MKDEBUG 2012-06-06 14:15:05 -03:00
Brian Fraser fraserb@gmail.com
3babaab15a Merged lp:~percona-toolkit-dev/percona-toolkit/pt-osc-cleanup-guards 2012-06-06 11:50:00 -03:00
Daniel Nichter
48677eb4d1 All Test::More::diag() in Sandbox.pm instead of just diag(). 2012-06-05 19:56:46 -07:00
Daniel Nichter
0e4b01bad7 Make wipe_clean() wait until 12347 is clean. Fix ChangeHandler.t, CompareResults.t, and Quoter.t so they leave the dbs clean. 2012-06-05 19:55:03 -07:00
Daniel Nichter
2d7cc5e8a4 Check for leftover databases. Fix MasterSlave.t so it doesn't leave leftover dbs and it resets 12347. 2012-06-05 18:55:32 -07:00
Brian Fraser fraserb@gmail.com
e57b67f10e Cxn & pt-osc: Guard the cleanup tasks. 
Does this twofold: By adding a __DIE__ hook to make sure they are
called, and by checking that the dbh is an actual database handle.
 2012-06-05 16:24:32 -03:00
Daniel Nichter
ac620957c9 Print on which host tables are different. 2012-06-05 11:20:18 -07:00
Daniel Nichter
a44a273800 Revert back to ok() for Sandbox::ok() so tests file don't have to be updated. Shorten verify_test_data_integrity() to verify_test_data(). 2012-06-05 11:02:48 -07:00
Daniel Nichter
b96e43419f Merge ~percona-toolkit-dev/percona-toolkit/fix-ptc-checksum-tbl-bug-1008778. 2012-06-05 13:08:21 -04:00
Daniel Nichter
4f83cf4102 Wait for --replicate table to repl on all slaves. Add wait_for arg to PerconaTest::full_output(). 2012-06-05 12:29:26 -04:00
Brian Fraser fraserb@gmail.com
345a21a82e Replace the last instances of MKDEBUG 2012-06-05 12:28:36 -03:00
Brian Fraser fraserb@gmail.com
6b93d51e43 ReadKeyMini: Restore STDIN flags in cooked() 
This will probably resolve bug #946707: pt-sift loses STDIN after pt-diskstats
 2012-06-05 12:16:49 -03:00
baron
95f7e53feb simplify and clean up sandbox/test-env a little, also helps prevent error messages like this when running test-env: [Warning] option 'thread_stack': unsigned value 126976 adjusted to 131072 2012-06-05 10:22:04 -04:00
Baron Schwartz
5c23b95072 magic to check that tests didn't muck with test dataset 2012-06-05 03:24:59 -04:00
Brian Fraser fraserb@gmail.com
6ede613752 VersionParser: Allow laxer version numbers, introduce the other cmp methods. 2012-06-05 13:50:41 -03:00
Baron Schwartz
2d33ba23dc Attempt to avoid race condition in t/pt-table-checksum/progress.t by slowing down the checksum job slightly, but this is a TEST_WISHLIST PLUGIN_WISHLIST and this test goes onto the spreadsheet of unreliable tests. 2012-06-05 00:09:31 -04:00
Daniel Nichter
1b11b7d7d9 Kill cxn only if using a db in Sandbox::wipe_clean(). Fix pt-duplicate-key-checker/basics.t for 5.0. Make pt-table-sync/lock_and_rename.t more stable. 2012-06-04 18:46:35 -07:00
Baron Schwartz
3a717b645c Sometimes Sandbox.pm tries to kill a connection that's already gone, and that kills the whole test. 2012-06-04 12:08:50 -04:00
Brian Fraser fraserb@gmail.com
9aaafe8478 PerconaTest: Stop no_diff from mistakenly unlinking samples 2012-06-04 21:57:58 -03:00
Brian Fraser fraserb@gmail.com
0c32f69ee0 PerconaTest: Split slurp_file out of load_file, also changes to no_diff. 
The no_diff changes introduce two new transformation options,
and also add a diag with the diff if there was one.
 2012-06-04 21:47:18 -03:00
Baron Schwartz
0637b85bb3 only commit if not autocommit; fixes some newly broken tests 2012-06-04 10:08:06 -04:00
Baron Schwartz
a074ce0526 Make pt-index-usage commit after saving results to tables, which fixes tests in MySQL 5.5 2012-06-04 08:21:02 -04:00
Brian Fraser fraserb@gmail.com
a3302dde93 Merge the fix for systems that lack sysctl 2012-06-04 12:01:12 -03:00
Brian Fraser fraserb@gmail.com
865b3e4e9d collect_system_info shouldn't assume that sysctl is available. 
It isn't on CentOS 5.5. It didn't really assume that it
was always there, but rather than the produced samples file
would be.
 2012-06-04 11:57:44 -03:00
Daniel Nichter
4e90f65736 Make pt-table-checksum/relication_filters.t and progress.t stable. Add PerconaTest::wait_until_no_lag(). Fix sandbox/test-env reset. 2012-06-04 07:33:14 -07:00
Daniel Nichter
7f51c42493 Fix Sandbox::leftover_servers(). Temporarily disable very slow test in pt-table-checksum/chunk_size.t. Debugging pt-table-checksum/replication_filters.t. 2012-06-04 08:24:30 -04:00
Baron Schwartz
48f642551e fix some problems that happen when there are many mysqld instances 2012-06-03 20:53:57 -04:00
Daniel Nichter
2afbe4a729 Handle NAME_lc in Sandbox::wipe_clean(). 2012-06-03 15:04:50 -07:00
Daniel Nichter
e8d9781123 Merge changes. 2012-06-03 14:51:03 -07:00
Daniel Nichter
696a3b5e68 Fix mem/file size shortening in pt-summary and pt-mysql-summary. 2012-06-03 14:49:35 -07:00
Baron Schwartz
6cc17a73c2 A few miscellaneous changes, including a fix for metadata locking and wipe_clean in Sandbox.pm 2012-06-03 17:20:12 -04:00
Daniel Nichter
98b384d845 Make pt-stalk/pt-stalk.t stable. Add PerconaTest::not_running(). 2012-06-03 13:29:45 -07:00
Baron Schwartz
87c8240369 fix problems with sandbox breaking replication and pt-slave-delay 2012-06-03 15:29:35 -04:00
Daniel Nichter
275ae93084 Fix waiting for slave message. 2012-06-03 14:41:29 -04:00
Daniel Nichter
67757bf838 Return false if the slave isn't ok. Use diag() for the waiting for slave message. 2012-06-03 14:21:29 -04:00
Daniel Nichter
a0b54b145d Merge 2.1. 2012-06-03 13:29:02 -04:00
Daniel Nichter
e5300c78a0 Check slave threads and read-only in Sandbox::ok(). 2012-06-03 13:09:46 -04:00
Daniel Nichter
d9a786d8f3 Add Sandbox::ok(). 2012-06-03 12:46:12 -04:00
Daniel Nichter
09ddcd641e Merge advisor-json-output and clean up pt-query-advisor docs. 2012-06-01 09:13:11 -06:00
Brian Fraser fraserb@gmail.com
0b5196a1f1 Merged ~percona-toolkit-dev/percona-toolkit/fix-1003315 2012-05-30 15:52:21 -03:00
Brian Fraser fraserb@gmail.com
8e04e3df60 Fix for 1003315: dry-run + alter-fk-method = auto always fail. 2012-05-30 14:27:07 -03:00
Brian Fraser fraserb@gmail.com
20d6ccff9e Fix for 923896 2012-05-27 23:28:35 -03:00
Daniel Nichter
fcd8958fb7 Validate --max|critical-load early and strongly, and check that the var exists and is defined. 2012-05-25 15:27:23 -06:00