lp:~akopytov/percona-xtrabackup/bug907147-1.6

Branch merges

Propose for merging

Merged into lp:percona-xtrabackup/1.6 at revision 313

Percona developers: Pending requested 2011-12-21

Diff: 31 lines (+15/-1)

2 files modified

innobackupex (+1/-1)
test/t/ib_binlog_info.sh (+14/-0)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Bug #907147: 1.6.4-313 loses master log info

Undecided

Fix Released

Link a bug report

Related blueprints

313. By Alexey Kopytov on 2011-12-21

Bug #907147: 1.6.4-313 loses master log info

The problem was that write_binlog_info() was filtering results coming
from keep-alive queries based in specific string. Which was changed
in the fix for bug #408803.

312. By Alexey Kopytov on 2011-12-17

Automerge lp:~lachlan-mulcahy/percona-xtrabackup/1.6-bug903984

311. By Stewart Smith on 2011-12-14

merge stewart's fix for /usr/bin/perl to /usr/bin/env perl

310. By Stewart Smith on 2011-12-14

merge Alexey's fix for Bug #514068: Output to STDOUT and STDERR is not conventional

309. By Alexey Kopytov on 2011-12-02

Merged from trunk.

308. By Alexey Kopytov on 2011-11-28

Manual merge from the bug branch (conflict in innobackupex).

307. By Alexey Kopytov on 2011-11-28

Merge from the BP tree.

306. By Alexey Kopytov on 2011-11-28

Merge from the BP tree.

305. By Alexey Kopytov on 2011-11-11

Bug #408803: innodbbackupex reaches Servers wait_timeout

The problem was the flawed keep-alive logic in innobackupex: keep-alive
queries were only sent to server when mysql_keep_alive() or
mysql_check() were explicitly called. Which led to connection timeouts
when some activity between those calls (e.g. copying a large non-InnoDB
file) took longer than server's wait_timeout.

Fixed innobackupex so that keep-alive queries are now sent automatically
using alarms. The time interval was also reduced from 1800 to 5 seconds
to accommodate low wait_timeout values. Since there is no need for
explicit calls to mysql_check() and mysql_keep_alive(), those functions
were removed.

There is no test case, because time sensitive tests for innobackupex are
basically impossible.

304. By Alexey Kopytov on 2011-11-01

Bug #884737: xtrabackup's --parallel option asserts / crashes with a
             value of -1

The problem was that my_getopt was implicitly converting negative
values to large positive ones, so xtrabackup would fail with an
asserts / crashes due to resource shortage.

Fixed by changing 'parallel' to a signed option so negative values are
detected and adjusted to the default value of 1 by my_getopt.