View Bazaar branches
Get this repository:
git clone https://git.launchpad.net/maria

MariaDB has 18 active reviews.

Branches

Name Last Modified Last Commit
bb-10.3-vicentiu-dtoa 2018-01-09 07:05:48 UTC 2018-01-09
Workaround for dtoa.c

Author: Vicențiu Ciorbaru
Author Date: 2018-01-09 07:05:48 UTC

Workaround for dtoa.c

There is no obvious reason why multiplying by 1.0 helps solve a
precision problem on gcc-7 or clang-5, but it does work. Casting to
ulong however makes the problem come up again. Windows warns about loss
of precision, so to quiet warnings until an appropriate fix is
implemented, ifdef 2 versions of the code.

bb-10.3-vicentiu-no-dtoa 2018-01-08 23:22:17 UTC 2018-01-08
Fixed mdev-14880: Assertion `inj_cond_list.elements' failed

Author: Igor Babaev
Author Date: 2018-01-08 23:21:52 UTC

Fixed mdev-14880: Assertion `inj_cond_list.elements' failed
in JOIN::inject_best_splitting_cond

The value of SplM_opt_info::last_plan should be set to NULL
before any search for a splitting plan for a splittable
materialized table.

bb-10.2-andrei-ext 2018-01-04 00:11:01 UTC 2018-01-04
The slave io thread does not conduct integrity check for row-based

Author: Andrei Elkin
Author Date: 2018-01-04 00:11:01 UTC

The slave io thread does not conduct integrity check for row-based
events. Specifically it tolerates missed STMT_END flag marked terminal
event that conclude a sequence of "elementary" "block" Rows-events.
Failure to react on a potentially missed event can confuse the applier thread
in various ways.

The slave io is made in this patch to track the STMT_END status.
Whenever at a first event following a sequence of Rows events the IO thread
finds out that a preceding Rows event did not actually had the flag, an
explicit error is issued.

Replication can be resumed as a provided test demonstrates.

Note that currently the row-based group integrity check is limited
to the version 1 rows event which are generated by Mariadb masters.

bb-10.3-igor 2017-12-31 00:19:17 UTC 2017-12-31
Merge branch '10.3' into bb-10.3-igor

Author: Igor Babaev
Author Date: 2017-12-31 00:19:17 UTC

Merge branch '10.3' into bb-10.3-igor

bb-10.3-vicentiu-travis 2017-12-30 15:18:24 UTC 2017-12-30
Tentative fix for travis

Author: Vicențiu Ciorbaru
Author Date: 2017-12-30 15:18:24 UTC

Tentative fix for travis

This undoes a lock-unlock for LOCK_thread_count.

bb-10.3-semisync 2017-12-18 11:43:38 UTC 2017-12-18
MDEV-13073. This patch is a followup of the previous one to convert the trail...

Author: Andrei Elkin
Author Date: 2017-11-27 19:06:17 UTC

MDEV-13073. This patch is a followup of the previous one to convert the trailing underscore identifier to mariadb standard. For identifier representing class private members the underscore is replaced with a `m_` prefix. Otherwise `_` is just removed.

bb-10.1-MDEV-14524 2017-12-13 10:52:53 UTC 2017-12-13
MDEV-14524 TokuDB is unable to be built on Linux

Author: Oleksandr Byelkin
Author Date: 2017-12-13 10:52:53 UTC

MDEV-14524 TokuDB is unable to be built on Linux

tokudb needs either F_NOCACHE or O_DIRECT, not both

bb-10.3-spider 2017-12-02 19:50:08 UTC 2017-12-02
Remove compiler warnings

Author: Monty
Author Date: 2017-12-02 19:50:08 UTC

Remove compiler warnings

bb-10.2-semisync_ali 2017-11-28 20:39:37 UTC 2017-11-28
MDEV-13073. This patch is a followup of the previous one to convert the trail...

Author: Andrei Elkin
Author Date: 2017-11-27 19:06:17 UTC

MDEV-13073. This patch is a followup of the previous one to convert the trailing underscore identifier to mariadb standard. For identifier representing class private members the underscore is replaced with a `m_` prefix. Otherwise `_` is just removed.

bb-10.2-semisync 2017-11-16 19:14:59 UTC 2017-11-16
MDEV-13073. Fixed failing on BB sys_vars.rpl_semi_sync_master_enabled_basic. ...

Author: Andrei Elkin
Author Date: 2017-11-16 19:14:59 UTC

MDEV-13073. Fixed failing on BB sys_vars.rpl_semi_sync_master_enabled_basic. The issue forced to add up aan argument to THD::THD and THD::init to propagate skip of a mutex locking. The lock could not be released instead as this leads to clear deadlock; the test is received a piece for concurrent ack thread start/stop.

howard0su-10.2 2017-11-16 08:39:48 UTC 2017-11-16
Support CRC32 SSE2 implementation under Windows

Author: Jun Su
Author Date: 2017-10-25 09:41:23 UTC

Support CRC32 SSE2 implementation under Windows

mdev-13626 2017-11-14 22:22:09 UTC 2017-11-14
innodb tests 563

Author: Larysa Sherepa
Author Date: 2017-11-14 22:22:09 UTC

innodb tests 563

szworkshop 2017-11-13 06:05:34 UTC 2017-11-13
Implement File Descriptor Plugin

Author: Vicențiu Ciorbaru
Author Date: 2017-11-13 06:05:34 UTC

Implement File Descriptor Plugin

Made use of /proc filesystem to get list of open file descriptors.
This is implemented as a Information_Schema plugin table PROC_FDINFO.

bb-10.2-spider-extra-jacob 2017-11-10 21:38:09 UTC 2017-11-10
Adding direct update/delete to the server and to the partition engine.

Author: Jacob Mathew
Author Date: 2017-10-14 22:24:24 UTC

Adding direct update/delete to the server and to the partition engine.

Contains Spiral patches:
- Spiral Patch 006: 006_mariadb-10.2.0.direct_update_rows.diff MDEV-7704
- Spiral Patch 008: 008_mariadb-10.2.0.partition_direct_update.diff MDEV-7706
- Spiral Patch 010: 010_mariadb-10.2.0.direct_update_rows2.diff MDEV-7708
- Spiral Patch 011: 011_mariadb-10.2.0.aggregate.diff MDEV-7709
- Spiral Patch 027: 027_mariadb-10.2.0.force_bulk_update.diff MDEV-7724
- Spiral Patch 061: 061_mariadb-10.2.0.mariadb-10.1.8.diff MDEV-12870

- Add support for direct update and direct delete requests. A direct
  update/delete request handles all qualified rows in a single operation,
  rather than one row at a time.

- This commit has the following differences compared to the original patches:
  - Most of the parameters of the new functions are unnecessary. The
    unnecessary parameters have been removed.
  - Changed bit positions for new handler flags upon consideration of
    handler flags not needed by other Spiral patches and handler flags
    merged from MySQL.

bb-10.3-MDEV-9059 2017-11-07 16:39:29 UTC 2017-11-07
MDEV-9059: protocol: bundle first command with the authentication packet

Author: Oleksandr Byelkin
Author Date: 2017-11-06 10:23:32 UTC

MDEV-9059: protocol: bundle first command with the authentication packet

bb-10.0-galera-andrii 2017-11-06 09:51:57 UTC 2017-11-06
re-enable tests which succeeded altogether twice with: --par=6 --repeat=10

Author: Andrii Nikitin
Author Date: 2017-11-06 09:51:57 UTC

re-enable tests which succeeded altogether twice with: --par=6 --repeat=10

svoj-szworkshop 2017-10-31 11:14:29 UTC 2017-10-31
PROC_FDINFO plugin for workshop

Author: Sergey Vojtovich
Author Date: 2017-10-31 11:14:29 UTC

PROC_FDINFO plugin for workshop

mdev-13625 2017-10-31 09:20:00 UTC 2017-10-31
deleted wl6501/(they are relevant only from 10.2.2)

Author: Larysa Sherepa
Author Date: 2017-10-31 09:20:00 UTC

deleted wl6501/(they are relevant only from 10.2.2)

bb-10.1-mdev9510-andrei 2017-10-25 14:23:18 UTC 2017-10-25
MDEV-9510 Segmentation fault in binlog thread causes crash

Author: Andrei Elkin
Author Date: 2017-10-16 16:32:02 UTC

MDEV-9510 Segmentation fault in binlog thread causes crash

With combination of --log-bin and Galera the server may crash
reporting two characteristic stacks:

  /usr/sbin/mysqld(_ZN13MYSQL_BIN_LOG13mark_xid_doneEmb+0xc7)[0x7f182a8e2cb7]
  /usr/sbin/mysqld(binlog_background_thread+0x2b5)[0x7f182a8e3275]

or

  /usr/sbin/mysqld(_ZN13MYSQL_BIN_LOG21do_checkpoint_requestEm+0x9d)[0x7ff395b2dafd]
  /usr/sbin/mysqld(_ZN13MYSQL_BIN_LOG20checkpoint_and_purgeEm+0x11)[0x7ff395b2db91]
  /usr/sbin/mysqld(_ZN13MYSQL_BIN_LOG16rotate_and_purgeEb+0xc2)[0x7ff395b300b2]

The reason of the failure appears to be non-matching decrements for
  `xid_count_per_binlog::xid_count`
which can occur when a transaction is executed having its connection issued
`SET @@sql_log_bin=0`. In such case the xid count is not incremented but
its decrements still runs to turn `binlog_xid_count_list` into improper state
which the following FLUSH BINARY LOGS exposes through the crash.

*Note_1*: the regression test reuses an existing galera.sql_log_bin
which does not run stably (even in its base form) by mtr with --log-bin.

*Note_2*: 10.0-galera branch is free of this issue having missed MDEV-7205
fixes.

svoj-workshop 2017-10-06 10:03:03 UTC 2017-10-06
Workshop plugin

Author: Sergey Vojtovich
Author Date: 2017-10-06 10:03:03 UTC

Workshop plugin

10.3-MDEV-11369 2017-10-06 06:51:14 UTC 2017-10-06
Merge 10.3 into 10.3-instant

Author: Marko Mäkelä
Author Date: 2017-10-06 06:51:14 UTC

Merge 10.3 into 10.3-instant

The tree state corresponds to the squashed
commit a4948dafcd7eee65f16d848bdc6562fc49ef8916 in the 10.3 branch.

This branch is preserved for future reference.

bb-10.2.4-spider-extra-jacob 2017-09-30 00:58:56 UTC 2017-09-30
Adding direct update/delete to the server and to the partition engine.

Author: Jacob Mathew
Author Date: 2017-09-20 22:05:34 UTC

Adding direct update/delete to the server and to the partition engine.

Contains Spiral patches:
- Spiral Patch 006: 006_mariadb-10.2.0.direct_update_rows.diff MDEV-7704
- Spiral Patch 008: 008_mariadb-10.2.0.partition_direct_update.diff MDEV-7706
- Spiral Patch 010: 010_mariadb-10.2.0.direct_update_rows2.diff MDEV-7708
- Spiral Patch 011: 011_mariadb-10.2.0.aggregate.diff MDEV-7709
- Spiral Patch 027: 027_mariadb-10.2.0.force_bulk_update.diff MDEV-7724
- Spiral Patch 061: 061_mariadb-10.2.0.mariadb-10.1.8.diff MDEV-12870

- Add support for direct update and direct delete requests. A direct
  update/delete request handles all qualified rows in a single operation,
  rather than one row at a time.

- This patch has the following differences compared to the original patch:
  - Most of the parameters of the new functions are unnecessary. The
    unnecessary parameters have been removed.
  - Changed bit positions for new handler flags upon consideration of
    handler flags not needed by other Spiral patches.

10.2-spider-extra 2017-09-29 14:48:32 UTC 2017-09-29
Fix index scan cleanup in the partition engine.

Author: Monty
Author Date: 2017-09-29 14:48:32 UTC

Fix index scan cleanup in the partition engine.

Spiral Patch 057: 057_mariadb-10.2.0.partition_index_end.diff MDEV-12999

Original author: Kentoku SHIBA
First reviewer: Jacob Mathew
Second reviewer: Michael Widenius

bb-10.2-vicentiu2 2017-09-19 12:13:53 UTC 2017-09-19
Merge remote-tracking branch 'connect/10.2' into 10.2

Author: Vicențiu Ciorbaru
Author Date: 2017-09-19 12:07:32 UTC

Merge remote-tracking branch 'connect/10.2' into 10.2

bb-10.2-mdev11934 2017-09-11 15:15:58 UTC 2017-09-11
MDEV-11934: MariaRocks: Group Commit with binlog

Author: Sergey Petrunia
Author Date: 2017-09-11 15:15:58 UTC

MDEV-11934: MariaRocks: Group Commit with binlog

Implement MariaDB's Group Commit API. This is a first attempt which lacks
the expected performance.

bb-10.0-mdev_13437-andrei 2017-08-28 13:55:23 UTC 2017-08-28
MDEV-13437 InnoDB: Failing assertion: !srv_read_only_mode in mtr0mtr.cc:322

Author: andrelkin
Author Date: 2017-08-28 13:55:23 UTC

MDEV-13437 InnoDB: Failing assertion: !srv_read_only_mode in mtr0mtr.cc:322

The assert was caused due to incorrect handling of --tc-heuristic-recover=ROLLBACK
and --innodb-read-only=1 option combination. Innodb attempted
rollback of a prepared transaction having no right to do so because of
the latter option.

Fixed with making Innodb to respect the read-only flag by
innobase_{commit,rollback}_by_xid (Marko's contribution).
LOCK_log mutex lock/unlock had to be added to cover MDEV-1438 "child" issue.
The measure is confirmed by mysql sources as well.

For testing of the conflicting option combination, mysql-test-run is
made to export a new $MYSQLD_LAST_CMD. It holds the very last value
generated by mtr.mysqld_start(). Even though the options have been
also always stored in $mysqld->{'started_opts'} there were no access
to them beyond the automatic server restart by mtr through the expect
file interface.

Effectively therefore $MYSQLD_LAST_CMD represents a more general interface to
$mysqld->{'started_opts'} which can be used in wider scopes including
server launch with incompatible options.

Notice another existing method to restart the server with incompatible
options relying on $MYSQLD_CMD is is aware of $mysqld->{'started_opts'} (the actual
options that the server is launched by mtr). In order to use this
method they would have to be provided manually.

Note, at merging to 10.2 an added search++ macro should be renamed
"back" to normal. It's being backparted to 10.0 "temporarily".

bb-10-ext-mdev-13369 2017-08-10 23:23:26 UTC 2017-08-10
Merge branch 'bb-10.2-ext' into bb-10-ext-mdev-13369

Author: Igor Babaev
Author Date: 2017-08-10 23:23:26 UTC

Merge branch 'bb-10.2-ext' into bb-10-ext-mdev-13369

# Conflicts:
# mysql-test/r/derived_cond_pushdown.result
# mysql-test/t/derived_cond_pushdown.test
# sql/sql_derived.cc
# sql/sql_select.cc
# sql/sql_select.h

bb-10.2-knielsen 2017-08-07 10:38:47 UTC 2017-08-07
MDEV-11937: InnoDB flushes redo log too often

Author: Kristian Nielsen
Author Date: 2017-08-07 10:38:47 UTC

MDEV-11937: InnoDB flushes redo log too often

Problem was introduced with the InnoDB 5.7 merge, the code related to
avoiding extra fsync at the end of commit when binlog is enabled. The
MariaDB method for this was removed, but the replacement MySQL method
based on thd_get_durability_property() is not functional in MariaDB.

This commit reverts the offending parts of the merge and adds a test
case, to fix the problem for InnoDB. But other storage engines are
likely to have a similar problem.

bb-10.2-mdev13153 2017-07-25 06:32:00 UTC 2017-07-25
A dummy change to force a buildbot re-run

Author: Sergey Petrunia
Author Date: 2017-07-25 06:32:00 UTC

A dummy change to force a buildbot re-run

bb-10.2-mdev13355 2017-07-21 17:20:14 UTC 2017-07-21
MDEV-13355: Assertion `using_unique_constraint || group_buff <= param->group_...

Author: Sergey Petrunia
Author Date: 2017-07-21 15:02:53 UTC

MDEV-13355: Assertion `using_unique_constraint || group_buff <= param->group_buff...

The code in OIN::optimize_inner() has this call

   calc_group_buffer(this, group_list)

the call is however bypassed when the optimizer figures out that the JOIN
has "Impossible WHERE".

If we attempt to calculate the value of a window function afterwards,
we will get a crash when trying to create a temporary table.

So, put a call to calc_group_buffer() here as well. It's a bit surprising
that all these steps are done for a query that will produce zero rows
but we are just following the approach taken by the fix for mdev-11999
here.

bb-10.2-mdev13352 2017-07-21 17:09:19 UTC 2017-07-21
MDEV-13351: Server crashes in st_select_lex::set_explain_type upon UNION with...

Author: Sergey Petrunia
Author Date: 2017-07-21 17:09:19 UTC

MDEV-13351: Server crashes in st_select_lex::set_explain_type upon UNION with window function

Make st_select_lex::set_explain_type() take into account that JOIN_TABs
it is traversing may be also post-join aggregation JOIN_TABs (which
have pos_in_table_list=NULL, etc).

bb-10.2-mdev13330 2017-07-16 08:28:15 UTC 2017-07-16
MDEV-13330: ANALYZE FORMAT=JSON should print time spent in SPs

Author: Sergey Petrunia
Author Date: 2017-07-16 08:28:15 UTC

MDEV-13330: ANALYZE FORMAT=JSON should print time spent in SPs

- More test coverage

10.2-cmake-submodules 2017-07-13 07:14:17 UTC 2017-07-13
cmake: update submodules automatically during the build

Author: Sergei Golubchik
Author Date: 2017-07-10 13:06:41 UTC

cmake: update submodules automatically during the build

bb-10.2.7 2017-07-12 18:35:37 UTC 2017-07-12
Rename mariadb-backup-10.2.files to mariadb-backup-10.2.install

Author: Vicențiu Ciorbaru
Author Date: 2017-07-12 18:31:14 UTC

Rename mariadb-backup-10.2.files to mariadb-backup-10.2.install

.files extension is not used by debian packaging, .install is.

bb-10.2-vicentiu3 2017-07-05 14:18:01 UTC 2017-07-05
MDEV-10879: Window Functions final ordering of result set

Author: Vicențiu Ciorbaru
Author Date: 2017-06-13 13:54:40 UTC

MDEV-10879: Window Functions final ordering of result set

Window functions now force an ordering to the result set if none is
specified through order by. The ordering that will be returned is the final ordering
that is used during window function computation. For multiple window functions,
it should generally be the most specific sort ordering for the final
window function in the select list.

bb-10.2-MDEV-12067 2017-07-03 06:48:07 UTC 2017-07-03
MDEV-12067 flashback does not correcly revert update/replace statements

Author: Lixun Peng
Author Date: 2017-07-03 06:48:07 UTC

MDEV-12067 flashback does not correcly revert update/replace statements

Problem
-------
For one-statement contains multiple row events, Flashback didn't reverse the
sequence of row events inside one-statement.

Solution
--------
Using a new array 'events_in_stmt' to store the row events of one-statement,
when parsed the last one event, then print from the last one to the first one.

In the same time, fixed another bug, without -vv will not insert the table_map
into print_event_info->m_table_map, then change_to_flashback_event() will not
execute because of Table_map_log_event is empty.

bb-10.2-serg-mariadb_config 2017-06-29 18:55:45 UTC 2017-06-29
MDEV-12938 Discrepancy between mysql_config and mariadb_config

Author: Sergei Golubchik
Author Date: 2017-06-29 17:21:40 UTC

MDEV-12938 Discrepancy between mysql_config and mariadb_config

bb-10.2-MDEV-7914 2017-06-29 01:42:20 UTC 2017-06-29
MDEV-7914 spider/bg.ha, spider/bg.ha_part fail sporadically in buildbot

Author: Jacob Mathew
Author Date: 2017-06-29 01:42:20 UTC

MDEV-7914 spider/bg.ha, spider/bg.ha_part fail sporadically in buildbot

Fixed the problem by adding a Spider shutdown indicator and a Spider memory
lock. Spider shutdown acquires the lock for write access and
all other requestors acquire the lock for read access.

bb-10.3-mdev12179 2017-06-26 15:29:30 UTC 2017-06-26
MDEV-12179: Per-engine mysql.gtid_slave_pos tables

Author: Sergey Petrunia
Author Date: 2017-06-26 15:29:30 UTC

MDEV-12179: Per-engine mysql.gtid_slave_pos tables

Merge to 10.3
- changed a few LEX_STRING to LEX_CSTRING

bb-10.1-for-elenst 2017-06-21 08:29:19 UTC 2017-06-21
Merge remote-tracking branch 'origin/10.0' into 10.1

Author: Vicențiu Ciorbaru
Author Date: 2017-06-21 08:29:19 UTC

Merge remote-tracking branch 'origin/10.0' into 10.1

For elenst only!

bb-10.2-spider-jacob 2017-06-16 21:25:10 UTC 2017-06-16
Spider Patch Integration

Author: Jacob Mathew
Author Date: 2017-06-16 21:25:10 UTC

Spider Patch Integration

Spiral patches 039 and 047.
Fixed bugs in patches 039 and 047, which resulted in a crash.

10.0-claudio 2017-05-31 14:46:13 UTC 2017-05-31
MDEV-10463: Granted as a whole to roles, databases are not show in SHOW DATAB...

Author: Vicențiu Ciorbaru
Author Date: 2017-05-22 10:38:26 UTC

MDEV-10463: Granted as a whole to roles, databases are not show in SHOW DATABASES

The problem lies in not checking role privileges as well during SHOW
DATABASES command. This problem is also apparent for SHOW CREATE
DATABASE command.

Other SHOW COMMANDS make use of check_access, which in turn makes use of
acl_get for both priv_user and priv_role parts, which allows them to
function correctly.

bb-10.0-MDEV-11958 2017-05-23 07:24:33 UTC 2017-05-23
c

Author: Oleksandr Byelkin
Author Date: 2017-05-23 07:24:33 UTC

c

bb-10.2-MDEV-10332 2017-05-09 12:54:00 UTC 2017-05-09
MDEV-10332 support for OpenSSL 1.1 and LibreSSL

Author: Sergei Golubchik
Author Date: 2017-05-03 19:22:59 UTC

MDEV-10332 support for OpenSSL 1.1 and LibreSSL

post-review fixes:
* move all ssl implementation related ifdefs/defines to one file
  (ssl_compat.h)
* work around OpenSSL-1.1 desire to malloc every EVP context by
  run-time checking that context allocated on the stack is big enough
  (openssl.c)
* use newer version of the AWS SDK for OpenSSL 1.1
* use get_dh2048() function as generated by openssl 1.1
  (viosslfactories.c)

bb-10.2-georg 2017-03-24 18:14:09 UTC 2017-03-24
Updated C/C tree

Author: Georg Richter
Author Date: 2017-03-24 18:14:09 UTC

Updated C/C tree

bb-10.2-MDEV-11419 2017-03-16 08:10:28 UTC 2017-03-16
MDEV-11419: Report all INSERT ID for bulk operation INSERT

Author: Oleksandr Byelkin
Author Date: 2016-11-29 20:49:25 UTC

MDEV-11419: Report all INSERT ID for bulk operation INSERT

Send all Insert IDs of the buld operation to client (JDBC need it)

bb-10.2-spider-MDEV-8954 2017-03-15 21:49:40 UTC 2017-03-15
MDEV-8954 unnecessary fetch of entire table

Author: Jacob Mathew
Author Date: 2017-03-15 21:49:40 UTC

MDEV-8954 unnecessary fetch of entire table

Avoid visiting each row multiple times during a SELECT DISTINCT query
on a Spider table.

MDEV-10332 2017-03-08 16:39:47 UTC 2017-03-08
Initial implementation for MDEV-10332:

Author: Georg Richter
Author Date: 2017-03-08 16:39:47 UTC

Initial implementation for MDEV-10332:
  support for OpenSSL 1.1 and LibreSSL
  tested against OpenSSL 1.0.1, 1.0.2, 1.1.0, Yassl and LibreSSL
  not working on Windows with native SChannel support, due to wrong cipher mapping: Latter one requires push of CONC-241 fixes.
  Please note that OpenSSL 0.9.8 and OpenSSL 1.1.0 will not work: Even if the build succeeds, test cases will fail with various errors, especially when using different tls libraries or versions for client and server.

MDEV-8954 2017-03-04 00:36:00 UTC 2017-03-04
Changes to fix Spider Performance Bug MDEV-8954 in 10.1

Author: Jacob Mathew
Author Date: 2017-03-04 00:36:00 UTC

Changes to fix Spider Performance Bug MDEV-8954 in 10.1

bb-10.1-wlad-xtrabackup-sst 2017-02-22 21:59:08 UTC 2017-02-22
Execute mariabackup in innobackupex mode to avoid "unknown argument" error.

Author: Nirbhay Choubey
Author Date: 2017-02-22 21:59:08 UTC

Execute mariabackup in innobackupex mode to avoid "unknown argument" error.

bb-10.2-mdev10694 2017-02-14 15:18:55 UTC 2017-02-14
MDEV-10694 - SIGFPE and/or huge memory allocation in maria_create ...

Author: Sergey Petrunia
Author Date: 2017-02-14 15:18:55 UTC

MDEV-10694 - SIGFPE and/or huge memory allocation in maria_create ...

The issue was that JOIN::rollup_write_data() used
JOIN::tmp_table_param::[start_]recinfo, which had uninitialized data.

These fields have uninitialized data, because JOIN::tmp_table_param
currently only stores some grouping-related data fields. The data about
the work (temporary) tables themselves is stored in
join->join_tab[...].tmp_table_param.

The fix is to make JOIN::rollup_write_data follow this convention
and look at the right TMP_TABLE_PARAM object

bb-10.2-mdev8112 2017-02-10 18:11:30 UTC 2017-02-10
[MDEV-8112] Port no slave left behind into 10.2

Author: Lixun Peng
Author Date: 2017-02-10 18:11:30 UTC

[MDEV-8112] Port no slave left behind into 10.2

This patch implements master throttling based on slave lag, aka no slave left behind.
The core feature works as follows:
1) The semi-sync-reply is ammended to also report back SQL-thread position (aka exec position)
2) Transactions are not removed from the "active-transaction-list"
   in the semi-sync-master plugin until atleast one slave has reported
   that it has executed this transaction. the slave lag can then
   be estimated by calculating how long the oldest transaction has been
   lingering in the active-transaction-list.
3) client-threads are forced to wait before commit until slave lag has decreased to acceptable value.

The following variables are introduced on master:
1. rpl_semi_sync_master_max_slave_lag (global)
2. rpl_semi_sync_master_slave_lag_wait_timeout (session)

The following status variables are introduced on master:
1. rpl_semi_sync_master_slave_lag_wait_sessions
2. rpl_semi_sync_master_estimated_slave_lag
3. rpl_semi_sync_master_trx_slave_lag_wait_time
4. rpl_semi_sync_master_trx_slave_lag_wait_num
5. rpl_semi_sync_master_avg_trx_slave_lag_wait_time

The following variables are introduced on slave:
1. rpl_semi_sync_slave_lag_enabled (global)

In addition to this, 2 optimizations that decreases overhead of semi-sync is introduced.
1) the idea of this is that if when a slave should send and transaction,
   it checks if it should be semi-synced, but rather
   than semi-sync:ing each transaction (which is done currently) the code
   will skip semi-syncing transaction if there already is newer transactions
   committed. But, since this can mean that semi-syncing is delayed indefinitely
   a cap is set using 2 new master variables:
   1. rpl_semi_sync_master_max_unacked_event_bytes (global)
   2. rpl_semi_sync_master_max_unacked_event_count (global)
2) rpl_semi_sync_master_group_commit which makes the semi-sync
   plugin only semi-sync the last transaction in a group commit.

bb-10.2-mdev7635 2017-02-10 01:51:29 UTC 2017-02-10
MDEV-7635: Fix for mroonga 32-bit test failures.

Author: Nirbhay Choubey
Author Date: 2017-02-10 01:42:05 UTC

MDEV-7635: Fix for mroonga 32-bit test failures.

bb-10.2-igor 2017-02-01 00:10:15 UTC 2017-02-01
Adjusted tests after the fix for bug mdev-9976.

Author: Igor Babaev
Author Date: 2017-02-01 00:10:15 UTC

Adjusted tests after the fix for bug mdev-9976.

Now ROWS is a reserved key word, so cannot be used
as an alias name.

bb-10.1-otto 2017-01-25 23:57:12 UTC 2017-01-25
Merge branch 'grooverdan-10.1-MDEV-11866-ANALYZE-FORMAT=JSON-volatility-norma...

Author: Sergey Petrunia
Author Date: 2017-01-25 23:57:12 UTC

Merge branch 'grooverdan-10.1-MDEV-11866-ANALYZE-FORMAT=JSON-volatility-normalise' into 10.1

bb-10.2-flashback 2017-01-24 16:55:37 UTC 2017-01-24
[MDEV-11871] mysqlbinlog test case fails caused by Flashback

Author: Lixun Peng
Author Date: 2017-01-24 16:55:37 UTC

[MDEV-11871] mysqlbinlog test case fails caused by Flashback

When mysqlbinlog parses event failed, then delete_dynamic(&binlog_events) will be ingored.
So it caused memory leak.

Moved delete_dynamic(&binlog_events) into cleanup().

MDEV-10570-plinux 2017-01-20 02:04:42 UTC 2017-01-20
[MDEV-10570] Merge Existing Community Patch “Flashback”

Author: Lixun Peng
Author Date: 2017-01-20 02:04:42 UTC

[MDEV-10570] Merge Existing Community Patch “Flashback”

==== Description ====

Flashback can rollback the instances/databases/tables to an old snapshot.
It's implement on Server-Level by full image format binary logs (--binlog-row-image=FULL), so it supports all engines.
Currently, it’s a feature inside mysqlbinlog tool (with --flashback arguments).

Because the flashback binlog events will store in the memory, you should check if there is enough memory in your machine.

==== New Arguments ====

--flashback (-B)
It will let mysqlbinlog to work on FLASHBACK mode.

==== Example ====

I have a table "t" in database "test", we can compare the output with "--flashback" and without.

#client/mysqlbinlog /data/mysqldata_10.0/binlog/mysql-bin.000001 -vv -d test -T t --start-datetime="2013-03-27 14:54:00" > /tmp/1.sql
#client/mysqlbinlog /data/mysqldata_10.0/binlog/mysql-bin.000001 -vv -d test -T t --start-datetime="2013-03-27 14:54:00" -B > /tmp/2.sql

Then, importing the output flashback file (/tmp/2.log), it can flashback your database/table to the special time (--start-datetime).
And if you know the exact postion, "--start-postion" is also works, mysqlbinlog will output the flashback logs that can flashback to "--start-postion" position.

==== Implement ====

1. As we know, if binlog_format is ROW (binlog-row-image=FULL in 10.1 and later), all columns value are store in the row event, so we can get the data before mis-operation.

2. Just do following things:

  2.1 Change Event Type, INSERT->DELETE, DELETE->INSERT.
  For example:
    INSERT INTO t VALUES (...) ---> DELETE FROM t WHERE ...
    DELETE FROM t ... ---> INSERT INTO t VALUES (...)

  2.2 For Update_Event, swapping the SET part and WHERE part.
  For example:
    UPDATE t SET cols1 = vals1 WHERE cols2 = vals2
    --->
    UPDATE t SET cols2 = vals2 WHERE cols1 = vals1

  2.3 For Multi-Rows Event, reverse the rows sequence, from the last row to the first row.
  For example:
    DELETE FROM t WHERE id=1; DELETE FROM t WHERE id=2; ...; DELETE FROM t WHERE id=n;
    --->
    DELETE FROM t WHERE id=n; ...; DELETE FROM t WHERE id=2; DELETE FROM t WHERE id=1;

  2.4 Output those events from the last one to the first one which mis-operation happened.
  For example:

bb-10.2-mdev7635.saved 2017-01-18 16:44:47 UTC 2017-01-18
MDEV-7635: Update tests to adapt to the new default sql_mode

Author: Nirbhay Choubey
Author Date: 2017-01-18 16:44:47 UTC

MDEV-7635: Update tests to adapt to the new default sql_mode

bb-10.1-nirbhay 2017-01-17 21:49:31 UTC 2017-01-17
MDEV-10954: sles11 does not have 'which' package (post-fix)

Author: Nirbhay Choubey
Author Date: 2017-01-17 21:49:31 UTC

MDEV-10954: sles11 does not have 'which' package (post-fix)

bb-10.2-cc 2017-01-17 11:17:00 UTC 2017-01-17
Updated C/C

Author: Georg Richter
Author Date: 2017-01-17 11:17:00 UTC

Updated C/C

bb-10.1-knielsen 2017-01-13 12:04:30 UTC 2017-01-13
Follow-up patch, update debian/patches/

Author: Kristian Nielsen
Author Date: 2017-01-13 12:04:30 UTC

Follow-up patch, update debian/patches/

bb-10.0-MDEV-10644 2016-12-29 07:31:12 UTC 2016-12-29
MDEV-10644 One of parallel replication threads remains active

Author: Lixun Peng
Author Date: 2016-12-29 07:31:12 UTC

MDEV-10644 One of parallel replication threads remains active
           after STOP SLAVE SQL_THREAD completes

Using mysql_mutex_trylock() instead of mysql_mutex_lock() for LOCK_active_mi in
function show_slave_received_heartbeats() and show_heartbeat_period().
If can't aquire LOCK_active_mi lock after 1 seconds, then give up and just show
SHOW_UNDEF.
This is a temporary fix. If want to solve related problems completely, should
change the LOCK_active_mi to mysql_rwlock_t.

bb-10.1-wsrep-ctas 2016-11-23 15:41:56 UTC 2016-11-23
Refs: MW-317: Post-fix

Author: Nirbhay Choubey
Author Date: 2016-11-23 15:41:56 UTC

Refs: MW-317: Post-fix

bb-10.2-nirbhay 2016-11-21 22:14:14 UTC 2016-11-21
MDEV-10792: Assertion `thd->mdl_context.is_lock_owner ..

Author: Nirbhay Choubey
Author Date: 2016-11-21 22:14:14 UTC

MDEV-10792: Assertion `thd->mdl_context.is_lock_owner ..

.. (MDL_key::TABLE, table->db, table->table_name, MDL_SHARED)'
failed in mysql_rm_table_no_locks

Reset error flag after temporary table has been
successfully dropped.

bb-10.0-MDEV-10863 2016-11-17 10:32:13 UTC 2016-11-17
Fix incorrect reading of events from relaylog in parallel replication.

Author: Kristian Nielsen
Author Date: 2016-11-16 10:00:38 UTC

Fix incorrect reading of events from relaylog in parallel replication.

The SQL thread keeps track of the position in the current relay log from
which to read the next event. This position is not normally used, but a
certain interaction with the IO thread can cause the SQL thread to re-open
the relay log and seek to the stored position.

In parallel replication, there were a couple of places where the position
was not updated. This created a race where a re-open of the relay log could
seek to the wrong position and start re-reading and processing events
already handled once, causing various kinds of problems.

Fix this by moving the position update into a single place in
apply_event_and_update_pos(), which should ensure that the position is
always updated in the parallel replication case.

This problem was found from the testcase of MDEV-10863, but it is logically
a separate problem.

bb-MDEV-10863 2016-11-16 10:11:00 UTC 2016-11-16
added a dash in front of extra version number

Author: Rasmus Johansson
Author Date: 2016-11-16 10:11:00 UTC

added a dash in front of extra version number

bb-10.2-mdev10953_1 2016-10-25 13:04:45 UTC 2016-10-25
remove sha2 from embedded build

Author: Georg Richter
Author Date: 2016-10-25 13:04:45 UTC

remove sha2 from embedded build

10.2-good_bye_yassl 2016-10-13 10:07:14 UTC 2016-10-13
- Replaced yassl by gnutls

Author: Georg Richter
Author Date: 2016-10-13 10:07:14 UTC

- Replaced yassl by gnutls
  TLS functionality is now handled in ma_tls_vio.c
- added new parameter ssl_passphrase for passphrase protected
  keys

bb-10.2-vicentiu-window 2016-09-23 15:15:01 UTC 2016-09-23
Make win_big test specify only if Sort_merge_passes have happened

Author: Vicențiu Ciorbaru
Author Date: 2016-09-23 15:15:01 UTC

Make win_big test specify only if Sort_merge_passes have happened

The number is irrelevant and may differ by architecture. We are only
interested if it happened or not

bb-10.2-mdev9736 2016-09-23 11:21:07 UTC 2016-09-23
MDEV-9736: Window functions: multiple cursors to read filesort result

Author: Sergey Petrunia
Author Date: 2016-09-23 11:18:29 UTC

MDEV-9736: Window functions: multiple cursors to read filesort result

Add support for having multiple IO_CACHEs with type=READ_CACHE to share
the file they are reading from.
Each IO_CACHE keeps its own in-memory buffer. When doing a read or seek
operation on the file, it notifies other IO_CACHEs that the file position
has been changed.

Make Rowid_seq_cursor use cloned IO_CACHE when reading filesort result.

bb-10.2-mdev9864 2016-09-22 08:45:05 UTC 2016-09-22
Allowed to use WITH clauses before SELECT in CREATE ... SELECT

Author: Igor Babaev
Author Date: 2016-09-22 08:45:05 UTC

Allowed to use WITH clauses before SELECT in CREATE ... SELECT
and INSERT ... SELECT.
Added test cases.

bb-10.2-decimal 2016-06-23 18:57:53 UTC 2016-06-23
Fixed compiler warnings and test failures found by buildbot

Author: Monty
Author Date: 2016-06-23 18:57:53 UTC

Fixed compiler warnings and test failures found by buildbot
Fixed ccfilter to detect errors where the column is included in the error message

bb-10.0.22-mdev8989 2016-06-15 14:23:47 UTC 2016-06-15
Backport of commit 904027c820fb344343de6f68a32bcc889f2bd407 to 10.0.21

Author: Sergey Petrunia
Author Date: 2016-06-15 14:23:47 UTC

Backport of commit 904027c820fb344343de6f68a32bcc889f2bd407 to 10.0.21

Commit message was:

  Merge branch 'bb-10.1-mdev8989' into 10.1

Some test results are probably not yet updated

MDEV-8947 2016-06-08 13:16:39 UTC 2016-06-08
Merge branch 'mdev9991' into mdev9991-10.1

Author: Kristian Nielsen
Author Date: 2016-06-08 13:16:39 UTC

Merge branch 'mdev9991' into mdev9991-10.1

bb-10.1-mdev8989 2016-06-02 14:31:56 UTC 2016-06-02
Update test results part #3.

Author: Sergey Petrunia
Author Date: 2016-06-02 14:31:56 UTC

Update test results part #3.

bb-10.1-elenst-mdev9909 2016-04-19 21:58:13 UTC 2016-04-19
Next part of experimental change for MDEV-9909

Author: Elena Stepanova
Author Date: 2016-04-19 21:58:13 UTC

Next part of experimental change for MDEV-9909

Do not allow file_key_management and example_key_management
to be loaded together, we already know it does not work (MDEV-9948)

bb-10.2-mdev8789 2016-02-17 22:30:25 UTC 2016-02-17
Addressed the issues raised in the review for the main patch

Author: Igor Babaev
Author Date: 2016-02-17 22:30:25 UTC

Addressed the issues raised in the review for the main patch
of mdev-8789.
Fixed a bug in TABLE_LIST::print.
Fixed another bug for the case when the definition of a
WITH table contained column list while the join in the main
query used two instances of this table.

10.0-tokudb-merge 2016-02-17 12:43:26 UTC 2016-02-17
TokuDB fixes to compile in MariaDB

Author: Sergei Golubchik
Author Date: 2016-02-17 12:43:26 UTC

TokuDB fixes to compile in MariaDB

bb-10.2-mdev8646 2016-02-09 20:35:59 UTC 2016-02-09
This is the consolidated patch for mdev-8646:

Author: Igor Babaev
Author Date: 2016-02-09 20:35:59 UTC

This is the consolidated patch for mdev-8646:
"Re-factor the code for post-join operations".

The patch mainly contains the code ported from mysql-5.6 and
created for two essential architectural changes:
1. WL#5558: Resolve ORDER BY execution method at the optimization stage
2. WL#6071: Inline tmp tables into the nested loops algorithm

The first task was implemented for mysql-5.6 by Ole John Aske.
It allows to make all decisions on ORDER BY operation at the optimization
stage.

The second task implemented for mysql-5.6 by Evgeny Potemkin adds JOIN_TAB
nodes for post-join operations that require temporary tables. It allows
to execute these operations within the nested loops algorithm that used to
be used before this task only for join queries. Besides these task moves
all planning on the execution of these operations from the execution phase
to the optimization phase.

Some other re-factoring changes of mysql-5.6 were pulled in, mainly because
it was easier to pull them in than roll them back. In particular all
changes concerning Ref_ptr_array were incorporated.

The port required some changes in the MariaDB code that concerned the
functionality of EXPLAIN and ANALYZE. This was done mainly by Sergey
Petrunia.

bb-10.1-mdev8646 2016-01-27 21:18:02 UTC 2016-01-27
Cleanup.

Author: Igor Babaev
Author Date: 2016-01-27 21:18:02 UTC

Cleanup.

10.2-travis-ci 2015-11-23 15:24:11 UTC 2015-11-23
Limit testsuite scope in Debian builds to make them pass faster

Author: Otto Kekäläinen
Author Date: 2015-11-11 05:20:48 UTC

Limit testsuite scope in Debian builds to make them pass faster

bb-10.0-knielsen 2015-11-13 14:30:48 UTC 2015-11-13
Merge branch 'mdev7818-4' into bb-10.0-knielsen

Author: Kristian Nielsen
Author Date: 2015-11-13 14:30:48 UTC

Merge branch 'mdev7818-4' into bb-10.0-knielsen

bb-10.1-default 2015-08-17 22:00:28 UTC 2015-08-17
Fixed failing tests and compiler warnings

Author: Monty
Author Date: 2015-08-17 22:00:28 UTC

Fixed failing tests and compiler warnings
- UNINIT_VAR() was required for 4.8.3 on openSUSE 13.2
- repair_symlink changed because of some of my earlier fixes for REPAIR

10.0-custombld 2015-06-29 09:52:03 UTC 2015-06-29
MDEV-8302: Duplicate key with parallel replication

Author: Kristian Nielsen
Author Date: 2015-06-29 09:41:06 UTC

MDEV-8302: Duplicate key with parallel replication

Intermediate patch to try to track down the root cause of the problem by
running a custom binary on the user's actual load.

This patch adds some printouts in the error log, all prefixed with
"MDEV8302:", that will hopefully trigger when the error condition occurs and
give more information about what is happening.

In addition, this patch implements what might or might not be a fix for the
error (along with a printout to show if the fix actually triggers).

The fix is to add a check for thd->killed just before a transaction does
mark_start_commit(). This could help reduce the chance of a transaction
detecting a deadlock kill only after running mark_start_commit(), which
might be a source of the original duplicate key problem (though no concrete
sequence of events is currently known to be able to cause this condition).

Finally, this patch marks the generated binaries with a -mdev8302a version
postfix to help identify them.

bb-10.1-mdev6115 2015-06-11 17:25:57 UTC 2015-06-11
A stub performing name resolution of window functions in simplest queries.

Author: Igor Babaev
Author Date: 2015-06-11 17:25:57 UTC

A stub performing name resolution of window functions in simplest queries.

bb-5.5-knielsen 2015-02-23 12:27:51 UTC 2015-02-23
MDEV-7310: last_commit_pos_offset set to wrong value after binlog rotate in g...

Author: Kristian Nielsen
Author Date: 2015-02-23 12:27:51 UTC

MDEV-7310: last_commit_pos_offset set to wrong value after binlog rotate in group commit

When the binlog was rotated due to @@max_binlog_size, the values of the
binlog_shapshot_file and binlog_snapshot_position were inconsistent in case of
non-transactional DML. The position was refering to the old file, while the
filename was of the new file after rotation. This patch makes them consistent
by making sure the position is also refering to the new file.

bb-10.1-mdev6657-r2 2014-10-07 08:40:10 UTC 2014-10-07
Merge branch '10.1' of github.com:MariaDB/server into 10.1

Author: Monty
Author Date: 2014-10-07 08:40:10 UTC

Merge branch '10.1' of github.com:MariaDB/server into 10.1

Merge with main branch

301387 of 387 results

Other repositories

Name Last Modified
lp:maria 1 hour ago
lp:~maria-captains/maria/+git/connector-c 2018-01-08
12 of 2 results
You can't create new repositories for MariaDB.