Skip to content

Tags: xenbo/mysql-server

Tags

mysql-5.5.56

Toggle mysql-5.5.56's commit message
Bug#25942414: SSL VARIABLES USAGE WHEN LIBMYSQL IS COMPILED WITH WITH…

…_SSL=NO

Description: If libmysql is compiled with WITH_SSL=NO,
             --ssl-* are not useful.

Solution: 1. Restricted WITH_SSL to values : bundled | yes | system
          2. Made "bundled" as default value for WITH_SSL. Also,
             not specifying WITH_SSL or even specifying WITH_SSL=no
             will be treated as/converted to WITH_SSL=bundled.

Reviewed-By: Tor Didriksen <[email protected]>
Reviewed-By: Georgi Kodinov <[email protected]>
(cherry picked from commit 3eb2058be34d1a21771fe89ff1a0c08f156899bc)

mysql-cluster-7.6.2

Toggle mysql-cluster-7.6.2's commit message
Merge branch 'mysql-5.7-cluster-7.5' into mysql-5.7-cluster-7.6

(cherry picked from commit b3cb4ea79732a1caab14d046d7642463da068a1d)

mysql-cluster-7.5.6

Toggle mysql-cluster-7.5.6's commit message
Bug#25654833 MYSQL CLUSTER 7.5.5 REPLICATION SLAVE SQL THREAD HANGS W…

…ITH CREATE TABLE

This is 'best effort' patch for the above problem, which seems to be a regression introduced in 7.5.5. Based on bug description, we *assume* this to be caused bu the 'part 3 of 4' push of patch for Bug#25042101 'SPLIT BINLOG INJECTOR_MUTEX IN TWO, DO REQUIRED CLEANUP':

.........
  Remove thread_yield() in binlog injector code previously put there
  as a temp stopgap in the commit below. This used to be required
  as the injector thread held the injector_mutex > 99% of the time when
  waiting for pollEvents(). That blocked client threads either wanting to
  access the data shared from the injector thread, or needing the injector_mutex
  while waiting for injector_cond to be signaled

  This should not be required anymore, as:

  1) injector_mutex has been splitt in two separate mutexes.
  2) We changed init of the injector_event_mutex from a 'FAST' to a 'SLOW'
     mutex which has better 'farness' properties in the scheduler
.........

The theory is that using the more 'fair' SLOW-mutex variant is not as 'fair' as assumed to be on all OS/VM variants. Thus the 'yields' may be unsafe to remove.

Bug has been hard to reproduce:

 - It has been 'randomly' reproduced using Oracle Linux 7 on top
   of 'Virtualbox'
 - It has than gone away for a couple of days.
 - It has *not* been reporoduced with this patch.

So it should be reasonable to expect that reintroducing the 'thread_yields' should solve the problem.

We cant say for sure though....

(cherry picked from commit f0e06cc2bc975ca5be084be9c307803e1da53cd4)

mysql-cluster-7.4.15

Toggle mysql-cluster-7.4.15's commit message
Add man page for ndb_print_frag_file for rpm

(cherry picked from commit c3d917f47f8b265b5a639245a498aa650187dac7)

mysql-cluster-7.2.28

Toggle mysql-cluster-7.2.28's commit message
Bug #25717383: MYSQLDUMP MAY EXECUTE ANY ARBITRARY QUERY

While writing comments if database object names has a new
line character, then next line is considered a command, rather
than a comment.
This patch fixes the way comments are constructed in mysqldump.

(cherry picked from commit 1099f9d17b1c697c2760f86556f5bae7d202b444)

mysql-cluster-7.3.17

Toggle mysql-cluster-7.3.17's commit message
Bug #25717383: MYSQLDUMP MAY EXECUTE ANY ARBITARY QUERY.

Post push fix.

mysql-8.0.1

Toggle mysql-8.0.1's commit message
Bug #25755137 post-push. Inadvertently dropped SET DEBUG_SYNC is rest…

…ored in the test. Otherwise there would be chances for the test to fail. Approved by Joao \thanks for discussion and offering an alternative approach.

mysql-5.7.18

Toggle mysql-5.7.18's commit message
Merge branch 'mysql-5.6' into mysql-5.7

(cherry picked from commit f1e6ace429bb95b417e46fa7e5104790dbca5412)

mysql-5.6.36

Toggle mysql-5.6.36's commit message
Bug #25717383: MYSQLDUMP MAY EXECUTE ANY ARBITARY QUERY.

Post push fix.

(cherry picked from commit de4b2f7f7c541b9e97df9a4c7b261065f1de7f6d)

mysql-5.5.55

Toggle mysql-5.5.55's commit message
Bug #25717383: MYSQLDUMP MAY EXECUTE ANY ARBITRARY QUERY

While writing comments if database object names has a new
line character, then next line is considered a command, rather
than a comment.
This patch fixes the way comments are constructed in mysqldump.

(cherry picked from commit 1099f9d17b1c697c2760f86556f5bae7d202b444)