Daniel Nichter
e2073065b1
EXPLAIN actual ascending nibble query to get MySQL's chosen index, and use that instead of the tool's chosen index. TODO: fix OobNibbleIterator.t and chunk_index.t.
2012-05-04 17:46:34 -06:00
Brian Fraser fraserb@gmail.com
85756b0e94
pt-kill: Add a --log-dsn option.
...
This is customer issue 23049.
2012-05-01 03:27:11 -03:00
Daniel Nichter
97b1209116
Remove useless var in SchemaIterator::_iterate_files() and update module in tools.
2012-04-03 12:21:15 -06:00
Daniel Nichter
c3b99d17a1
Add func to col ident struct to fix QueryAdvisorRules.t.
2012-04-03 11:03:17 -06:00
Daniel Nichter
0286631514
Merge summary-tools-2.1.
2012-04-03 10:14:55 -06:00
Daniel Nichter
b0b65fd8fd
Merge fix-log-parser-writer-bug-963225.
2012-04-03 09:56:11 -06:00
Brian Fraser fraserb@gmail.com
a01de98434
Fix for a bug in 32-bit Debian.
...
Namely, awk's printf("%d") doing what it's supposed to do, cast
it's operand to (int), which screws up large numbers.
The fix is simple, although it mostly just pushes the problem
under the rug; using %.0f instead of %d.
2012-04-03 12:50:12 -03:00
Daniel Nichter
0089f31a35
Merge pt-osc-2.1. Fix NibbleItertor.pm to work with OobNibbleIterator when resuming at oob boundaries (t/pt-table-checksum/resume.t was failing).
2012-04-03 09:33:10 -06:00
Brian Fraser fraserb@gmail.com
b807088b8e
Other instances of the substr(var, 0, n) bug
2012-04-03 11:29:36 -03:00
Brian Fraser fraserb@gmail.com
cfcef354b0
section(): Less printf, more awk
2012-04-03 11:07:30 -03:00
Brian Fraser fraserb@gmail.com
12bd26022a
Fix a section() bug.
...
Turns out that awk's indexes are documented to start from 1;
So what happens if you use 0? Apparently it's undefined behavior, so
substr(var, 0, 5);
and
substr(var, 1, 5);
May or may not actually do the same thing. In Debian 6, the former
doesn't quite work, and actually returns something like substr(var, 1, 4),
which broke section()'s output.
2012-04-03 10:55:51 -03:00
Daniel Nichter
403c93db3b
Handle Key_name or key_name in NibbleIterator::_get_index_cardinality().
2012-04-02 19:24:38 -06:00
Daniel Nichter
5eaa565566
Catch lost cleanup task (should rarely happen).
2012-04-02 19:20:56 -06:00
Daniel Nichter
34ada21344
Add comment about code that will fail with new Cxn due to NAME_lc.
2012-04-02 19:12:28 -06:00
Brian Fraser fraserb@gmail.com
c0dedd19bf
- section() revamp
...
- Redirect STDERR for the collection part in both tools
- rm_tempdir in sigtrap
- A part of pt-mysql-summary still depended on the system being run,
rather than the samples being read. Now it works properly.
2012-04-02 19:25:17 -03:00
Brian Fraser fraserb@gmail.com
6dbbd6de3b
The previous merge went a bit wrong, and left the libraries unsynced with the tools; This commit fixes it.
2012-04-02 13:36:17 -03:00
Daniel Nichter
c04b85a066
Update SchemaIterator, TableParser, and NibbleIterator in tools that use them. All tools' tests still pass.
2012-03-31 09:59:15 -06:00
Brian Fraser fraserb@gmail.com
761f8a5aac
Merge more of Baron's changes, also re-implement the MySQL Executables section
...
Plus other bugfixes.
2012-03-30 23:00:57 -03:00
Daniel Nichter
2bae569429
Don't use TABLE STATUS, use CREATE TABLE instead for the table's engine.
2012-03-30 18:46:08 -06:00
Daniel Nichter
1cd990a0f0
Update SchemaIterator in pt-table-sync. Make the module handle NAME_lc or not.
2012-03-30 18:15:54 -06:00
Brian Fraser fraserb@gmail.com
978b236985
Save the config file itself, not just its path.
2012-03-30 20:18:39 -03:00
Daniel Nichter
34602e89fe
Add pt-table-usage, copied and converted from mk-table-usage.
2012-03-30 16:45:10 -06:00
Daniel Nichter
4300f2908b
Add pt-fingerprint.
2012-03-30 16:07:37 -06:00
baron
06dc18ec46
clean up a couple of things causing tests to fail for me, and remove a feature that's not quite ready yet, and add another feature for innodb compression
2012-03-30 17:07:05 -04:00
baron
d44e8b539b
clean up a couple of things causing tests to fail for me, and remove a feature that's not quite ready yet, and add another feature for innodb compression
2012-03-30 17:04:25 -04:00
Brian Fraser fraserb@gmail.com
562a550a39
Fixes
2012-03-30 17:23:55 -03:00
Brian Fraser fraserb@gmail.com
b007016215
lib/Cxn.pm: Silence a warning (in 5.14)
...
I think this warning actually happens everywhere, but because of
how it happens, it gets masked by older versions of Perl.
Basically, Cxn's DESTROY method checks if $self->{dbh}, and if true,
calls ->disconnect on that. However, during global destruction --
when DESTROY is called -- you have no assurances that DESTROY
will be called before the dbh was already reaped.
This commit makes sure that the $self->{dbh} we have inside DESTROY
is actually a reference.
2012-03-30 16:03:04 -03:00
Brian Fraser fraserb@gmail.com
33954b72f9
Silence warnings from lib/QueryAdvisorRules.pm
2012-03-30 15:58:00 -03:00
Brian Fraser fraserb@gmail.com
8a4bc61a6b
Fix test failures in t/lib/SQLParser.t from an lc() on a variable without a default on lib/Schema.pm
2012-03-30 15:47:37 -03:00
Daniel Nichter
94d9483d03
Make vmstat, iostat, and mpstat run for --run-time not --interval.
2012-03-30 11:35:32 -06:00
Brian Fraser fraserb@gmail.com
33a4096663
Fix CPU_ARCH collection for Linux
2012-03-30 13:44:21 -03:00
Daniel Nichter
688570de79
Quote chunk col in char chunk sql.
2012-03-30 10:31:36 -06:00
baron@percona.com
26944c11cc
Update samples and tests for SlowLogWriter.pm and SlowLogParser.pm for bug 963225
2012-03-29 22:16:53 -04:00
Brian Fraser fraserb@gmail.com
f8113b4676
A parse_ethtools() that actually works
2012-03-29 19:44:51 -03:00
Brian Fraser fraserb@gmail.com
996a76a68e
pt-summary shouldn't leave empty files; both summary tools should
...
die if passed a --save-samples directory that has files in it.
2012-03-29 19:21:32 -03:00
Daniel Nichter
fcc3a44e23
Update alter_active_table.t and fix NibbleIterator to handle a growing table.
2012-03-29 15:23:38 -06:00
Brian Fraser fraserb@gmail.com
3ee8f9ba71
Fixed the Table Cache bugs.
...
Also makes t/lib/bash/collect_system_info.sh actually call setup_commands.
2012-03-28 15:29:57 -03:00
Brian Fraser fraserb@gmail.com
4bb5968c23
More fixes as per the reviews.
...
- Took out ${prefix}-
- get_var sanity
- pt-mysql-summary now follows pt-summary's suit and doesn't take
in files to write to
- pt-summary defers setting the CMD_FOO vars to a function, as
otherwise it would skip our PATH mungling.
2012-03-28 13:04:14 -03:00
Daniel Nichter
02b3574582
Add critical load to MySQLStatusWaiter.pm and pt-osc. Clean up the tool's docu.
2012-03-27 19:17:17 -06:00
Brian Fraser fraserb@gmail.com
ae6c1e3789
More fixes:
...
- pt-mysql-summary's --read-samples shouldn't need an active connection
- Bunch of pt-summary bugs, particularly several relating to disks.
2012-03-27 12:53:31 -03:00
Brian Fraser fraserb@gmail.com
0e4fa22d18
*-summary tools: Changes from the reviews. Tests for pt-mysql-summary not finished yet
2012-03-27 06:19:37 -03:00
Daniel Nichter
beaa9240e7
Rename preserve_embedded_numbers to match_embedded_numbers, and fingerprint_md5 to match_md5_checksums. Add corresponding options to pt-fingerprint.
2012-03-26 16:40:46 -06:00
Daniel Nichter
a81d25c489
Add fingerprint_md5 and preserve_embedded_numbers flags to QueryRewriter, used in fingerprint() to handle customer's requirements.
2012-03-26 10:45:46 -06:00
Daniel Nichter
45c1f3bfb6
Make NibbleIter query comments customizable. Update the module in pt-table-checksum and pt-osc.
2012-03-25 10:37:19 -06:00
Brian Fraser fraserb@gmail.com
3ffda5677b
pt-summary for 2.1, also changes to pt-mysql-summary from Daniel's review
2012-03-24 14:20:09 -03:00
Daniel Nichter
20d2aa7a08
Make NibbleIterator check for options before using them.
2012-03-22 11:28:58 -06:00
Brian Fraser fraserb@gmail.com
88a9a66946
pt-mysql-summary for 2.1
2012-03-20 18:29:35 -03:00
Brian Fraser fraserb@gmail.com
55c7248c54
(more syncing)
2012-03-20 17:16:06 -03:00
Daniel Nichter
a17bdd65d3
Rewriting pt-online-schema-change 2.1.1. Work in progress (this code doesn't work yet).
2012-03-20 13:34:16 -06:00
Brian Fraser fraserb@gmail.com
31afeb73b5
(temp commit, syncing up)
2012-03-20 11:21:16 -03:00
