Commit graph

355 commits

Author SHA1 Message Date
Vicențiu Ciorbaru
45898c2092 Merge remote-tracking branch 'origin/10.0' into 10.0 2017-05-18 15:45:55 +03:00
Marko Mäkelä
9f57e595b4 Refactor trx_undo_report_row_operation()
Fix a -fsanitizer=undefined warning that trx_undo_report_row_operation()
was being passed thr=NULL when the BTR_NO_UNDO_LOG_FLAG flag was set.

trx_undo_report_row_operation(): Remove the first two parameters.
The parameter clust_entry!=NULL distinguishes inserts from updates.

This should be a non-functional change (no observable change in
behaviour; slightly smaller code).
2017-05-18 09:06:52 +03:00
Marko Mäkelä
9f89b94ba6 MDEV-12358 Work around what looks like a bug in GCC 7.1.0
The parameter thr of the function btr_cur_optimistic_insert()
is not declared as nonnull, but GCC 7.1.0 with -O3 is wrongly
optimizing away the first part of the condition
UNIV_UNLIKELY(thr && thr_get_trx(thr)->fake_changes)
when the function is being called by row_merge_insert_index_tuples()
with thr==NULL.

The fake_changes is an XtraDB addition. This GCC bug only appears
to have an impact on XtraDB, not InnoDB.

We work around the problem by not attempting to dereference thr
when both BTR_NO_LOCKING_FLAG and BTR_NO_UNDO_LOG_FLAG are set
in the flags. Probably BTR_NO_LOCKING_FLAG alone should suffice.

btr_cur_optimistic_insert(), btr_cur_pessimistic_insert(),
btr_cur_pessimistic_update(): Correct comments that disagree with
usage and with nonnull attributes. No other parameter than thr can
actually be NULL.

row_ins_duplicate_error_in_clust(): Remove an unused parameter.

innobase_is_fake_change(): Unused function; remove.

ibuf_insert_low(), row_log_table_apply(), row_log_apply(),
row_undo_mod_clust_low():
Because we will be passing BTR_NO_LOCKING_FLAG | BTR_NO_UNDO_LOG_FLAG
in the flags, the trx->fake_changes flag will be treated as false,
which is the right thing to do at these low-level operations
(change buffer merge, ALTER TABLE…LOCK=NONE, or ROLLBACK).
This might be fixing actual XtraDB bugs.

Other callers that pass these two flags are also passing thr=NULL,
implying fake_changes=false. (Some callers in ROLLBACK are passing
BTR_NO_LOCKING_FLAG and a nonnull thr. In these callers, fake_changes
better be false, to avoid corruption.)
2017-05-17 16:09:22 +03:00
Vicențiu Ciorbaru
b87873b221 Merge branch 'merge-innodb-5.6' into bb-10.0-vicentiu
This merge reverts commit 6ca4f693c1ce472e2b1bf7392607c2d1124b4293
from current 5.6.36 innodb.

Bug #23481444	OPTIMISER CALL ROW_SEARCH_MVCC() AND READ THE
                       INDEX APPLIED BY UNCOMMITTED ROW
Problem:
========
row_search_for_mysql() does whole table traversal for range query
even though the end range is passed. Whole table traversal happens
when the record is not with in transaction read view.

Solution:
=========

Convert the innodb last record of page to mysql format and compare
with end range if the traversal of row_search_mvcc() exceeds 100,
no ICP involved. If it is out of range then InnoDB can avoid the
whole table traversal. Need to refactor the code little bit to
make it compile.

Reviewed-by: Jimmy Yang <jimmy.yang@oracle.com>
Reviewed-by: Knut Hatlen <knut.hatlen@oracle.com>
Reviewed-by: Dmitry Shulga <dmitry.shulga@oracle.com>
RB: 14660
2017-05-17 14:53:28 +03:00
Vicențiu Ciorbaru
0af9818240 5.6.36 2017-05-15 17:17:16 +03:00
Vicențiu Ciorbaru
1acfa942ed Merge branch '5.5' into 10.0 2017-03-03 01:37:54 +02:00
Jan Lindström
108b211ee2 Fix gcc 6.3.x compiler warnings.
These are caused by fact that functions are declared with
__attribute__((nonnull)) or left shit like ~0 << macro
when ~0U << macro should be used.
2017-02-16 12:02:31 +02:00
Sergei Golubchik
a79d46c3a4 Merge branch 'merge-innodb-5.6' into 10.0 2016-06-21 14:58:19 +02:00
Sergei Golubchik
720e04ff67 5.6.31 2016-06-21 14:21:03 +02:00
Sergei Golubchik
5b3c100788 Merge branch 'merge/merge-innodb-5.6' into 10.0 2015-12-13 10:18:42 +01:00
Sergei Golubchik
e9eaaa4b4a 5.6.28 2015-12-13 10:11:49 +01:00
Sergei Golubchik
04af573d65 Merge branch 'merge-innodb-5.6' into 10.0 2015-10-09 17:47:30 +02:00
Sergei Golubchik
86ff4da14d 5.6.27 2015-10-09 17:21:46 +02:00
Sergei Golubchik
82e9f6d948 Merge remote-tracking branch 'mysql/5.5' into 5.5 2015-10-08 22:54:24 +02:00
Thirunarayanan Balathandayuthapani
18b2c77780 Merge branch 'mysql-5.1' into mysql-5.5 2015-07-29 18:27:39 +05:30
Sergei Golubchik
6d06fbbd1d move to storage/innobase 2015-05-04 19:17:21 +02:00
Jan Lindström
a7fd11b31d MDEV-7685: MariaDB - server crashes when inserting more rows than
available space on disk

Add error handling when disk full situation happens and
intentionally bring server down with stacktrace because
on all cases InnoDB can't continue anyway.
2015-03-09 18:33:25 +02:00
Sergei Golubchik
6b05688f6d innodb 5.6.23 2015-02-18 17:59:21 +01:00
Jan Lindström
454beee5fb MDEV-6288 :Innodb causes server crash after disk full, then can't ALTER TABLE any more
Fix try to avoid unnecessary crashes when disk full situation is reached
on alter table.
2015-02-13 11:49:31 +02:00
Sergei Golubchik
476a8660e6 InnoDB 5.6.22 2015-01-21 14:33:39 +01:00
Jan Lindström
89a3628b0b Better comments part 2 with proof and simplified implementation.
Thanks to Daniel Black.
2014-11-25 12:04:32 +02:00
Jan Lindström
e3ded84b83 Fix typo. 2014-11-25 08:22:10 +02:00
Jan Lindström
e5802c38f9 Better comments and add a test case. 2014-11-25 08:06:41 +02:00
Jan Lindström
e15a83c0c7 Better comments part 2 with proof and simplified implementation.
Thanks to Daniel Black.
2014-11-26 16:41:28 +02:00
Jan Lindström
b62c4c6586 Better comments and add a test case. 2014-11-25 08:31:03 +02:00
Sergei Golubchik
a9a6bd5256 InnoDB 5.6.21 2014-11-20 16:59:22 +01:00
Sergei Golubchik
3c12c27907 5.5 merge 2014-11-20 16:07:34 +01:00
Jan Lindström
8bc5eabea8 MDEV-7084: innodb index stats inadequate using constant
innodb_stats_sample_pages

Analysis: If you set the number of analyzed pages 
to very low number compared to actual pages on 
that table/index it randomly pics those pages 
(default 8 pages), this leads to fact that query 
after analyze table returns different results. If 
the index tree is small, smaller than 10 * 
n_sample_pages + total_external_size, then the 
estimate is ok. For bigger index trees it is 
common that we do not see any borders between 
key values in the few pages we pick. But still 
there may be n_sample_pages different key values, 
or even more. And it just tries to 
approximate to n_sample_pages (8).

Fix: (1) Introduced new dynamic configuration variable
innodb_stats_sample_traditional  that retains
the current design. Default false.

(2) If traditional sample is not used we use
n_sample_pages = max(min(srv_stats_sample_pages,
                         index->stat_index_size),
                     log2(index->stat_index_size)*
                          srv_stats_sample_pages);

(3) Introduced new dynamic configuration variable
stat_modified_counter (default = 0) if set
sets lower bound for row updates when statistics is re-estimated.

If user has provided upper bound for how many rows needs to be updated
before we calculate new statistics we use minimum of provided value
and 1/16 of table every 16th round. If no upper bound is provided
(srv_stats_modified_counter = 0, default) then calculate new statistics
if 1 / 16 of table has been modified
since the last time a statistics batch was run.
We calculate statistics at most every 16th round, since we may have
a counter table which is very small and updated very often.
@param t table
@return true if the table has changed too much and stats need to be
recalculated
*/
#define DICT_TABLE_CHANGED_TOO_MUCH(t) \
	((ib_int64_t) (t)->stat_modified_counter > (srv_stats_modified_counter ? \
	ut_min(srv_stats_modified_counter, (16 + (t)->stat_n_rows / 16)) : \
		16 + (t)->stat_n_rows / 16))
2014-11-19 20:27:34 +02:00
Jan Lindström
3486135bb5 MDEV-6928: Add trx pointer to struct mtr_t
Merge Facebook commit 25295d003cb0c17aa8fb756523923c77250b3294
authored by Steaphan Greene from https://github.com/facebook/mysql-5.6

This adds a pointer to the trx to each mtr.
This allows the trx to be accessed in parts of the code
where it was otherwise not available. This is needed later.
2014-10-24 22:26:31 +03:00
Sergei Golubchik
75796d9ecb InnoDB 5.6.20 2014-09-11 16:42:54 +02:00
Michael Widenius
a1c1700b89 Fixed some compiler warnings 2014-07-30 10:05:01 +03:00
Sergei Golubchik
2023fac281 innodb-5.6.19 2014-08-06 20:05:10 +02:00
Jan Lindström
a3acd72570 Merge InnoDB fixes from 5.5 revisions 4229, 4230, 4233, 4237 and 4238 i.e.
4229: MDEV-5670: Assertion failure in file buf0lru.c line 2355
      Add more status information if repeatable.

4230: MDEV-5673: Crash while parallel dropping multiple tables under heavy load
      Improve long semaphore wait output to include all semaphore waits
      and try to find out if there is a sequence of waiters.

4233: Fix compiler errors on product build.

4237: Fix too agressive long semaphore wait output and add guard against introducing
      compression failures on insert buffer.

4238: Fix test failure caused by simulated compression failure on
      IBUF_DUMMY table.
2014-07-25 10:30:16 +03:00
Jan Lindström
67eb6f33a9 MDEV-5670: Assertion failure in file buf0lru.c line 2355
Add more status information if repeatable.
2014-07-22 22:08:06 +03:00
Sergei Golubchik
8ee9d19607 innodb 5.6.17 2014-05-07 17:32:23 +02:00
Sergei Golubchik
e2e5d07b28 MDEV-6184 10.0.11 merge
InnoDB 5.6.16
2014-05-06 09:57:39 +02:00
Jan Lindström
772aa0c575 MDEV-6160: InnoDB: Failing assertion: page_is_comp(next_page) == page_is_comp(page)
This is not yet a fix. This is change to print additional information at the point
when this assertion is going to happen. Print as much information about the pages
and index to find out why next page is not a compact format.
2014-04-28 08:30:05 +03:00
Sergei Golubchik
15ee97214d InnoDB 5.6.15 merge.
update test results
2014-02-26 19:36:33 +01:00
Sergei Golubchik
27d45e4696 MDEV-5574 Set AUTO_INCREMENT below max value of column.
Update InnoDB to 5.6.14
Apply MySQL-5.6 hack for MySQL Bug#16434374
Move Aria-only HA_RTREE_INDEX from my_base.h to maria_def.h (breaks an assert in InnoDB)
Fix InnoDB memory leak
2014-02-01 09:33:26 +01:00
Sergei Golubchik
37d240ecf9 MySQL-5.5.35 merge 2014-01-22 15:29:36 +01:00
Sergei Golubchik
d28d3ba40d 10.0-base merge 2013-12-16 13:02:21 +01:00
Sergei Golubchik
fa3f8a18b2 mysql-5.5.34 merge
(some patches reverted, test case added)
2013-11-19 13:16:25 +01:00
Mattias Jonsson
6d483cf3ab post-push fix for bug#17565888/bug#14621190 2013-10-28 17:08:31 +01:00
Mattias Jonsson
e8077731bd post-push fix for bug#17565888/bug#14621190 2013-10-28 17:08:31 +01:00
Mattias Jonsson
7a524cee32 post-push fix for bug17565888. 2013-10-19 15:29:51 +02:00
Mattias Jonsson
aceb173714 post-push fix for bug17565888. 2013-10-19 15:29:51 +02:00
Mattias Jonsson
20f579ea16 Bug#17565888: UP TO 75% INCREASE IN DBT3 QUERY (Q7) EXECUTION TIME
Regression from bug#14621190 due to disabled optimistic restoration
of cursor, which required full key lookup instead of verifying
if previously positioned btree cursor could be reused.

Fixed by enable optimistic restore and adjust cursor afterward.

rb#3324 approved by Marko.
2013-10-18 19:25:59 +02:00
Mattias Jonsson
f4fd68857b Bug#17565888: UP TO 75% INCREASE IN DBT3 QUERY (Q7) EXECUTION TIME
Regression from bug#14621190 due to disabled optimistic restoration
of cursor, which required full key lookup instead of verifying
if previously positioned btree cursor could be reused.

Fixed by enable optimistic restore and adjust cursor afterward.

rb#3324 approved by Marko.
2013-10-18 19:25:59 +02:00
Mattias Jonsson
6ca4be9c03 Bug#14621190: HA_INNOBASE::INDEX_NEXT SKIPS A RECORD IF PREVIOUS
INDEX_READ_MAP HAD NO MATCH

If index_read_map is called for exact search and no matching records
exists it will position the cursor on the next record, but still having the
relative position to BTR_PCUR_ON.
This will make a call for index_next to read yet another next record,
instead of returning the record the cursor points to.

Fixed by setting pcur->rel_pos = BTR_PCUR_BEFORE if an exact
[prefix] search is done, but failed.

Also avoids optimistic restoration if rel_pos != BTR_PCUR_ON,
since btr_cur may be different than old_rec.

rb#3324, approved by Marko and Jimmy
2013-10-01 10:05:45 +02:00
Mattias Jonsson
261268d84c Bug#14621190: HA_INNOBASE::INDEX_NEXT SKIPS A RECORD IF PREVIOUS
INDEX_READ_MAP HAD NO MATCH

If index_read_map is called for exact search and no matching records
exists it will position the cursor on the next record, but still having the
relative position to BTR_PCUR_ON.
This will make a call for index_next to read yet another next record,
instead of returning the record the cursor points to.

Fixed by setting pcur->rel_pos = BTR_PCUR_BEFORE if an exact
[prefix] search is done, but failed.

Also avoids optimistic restoration if rel_pos != BTR_PCUR_ON,
since btr_cur may be different than old_rec.

rb#3324, approved by Marko and Jimmy
2013-10-01 10:05:45 +02:00