The test case makes use of the fine DEBUG_SYNC facility. Furthermore,
since it needs synchronization on internal threads (dump and SQL
threads) the server code has DEBUG_SYNC commands internally deployed
and activated through the DBUG_EXECUTE_IF macro. The internal
DBUG_SYNC commands are then controlled from the test case through the
DEBUG variable.
There were three problems around the DEBUG + DEBUG_SYNC facility
usage:
1. When signaling the SQL thread to continue, the test would reset
immediately the DEBUG_SYNC variable. This could mean that the SQL
thread might loose the signal and continue to wait forever;
2. A similar scenario was happening with the dump thread on the
master. This thread was instructed to wait, and later it would be
signaled to continue, but immediately after the DEBUG_SYNC would be
reset. This could lead to the dump thread missing the signal and
wait forever;
3. The test was not cleaning itself up with respect to the
instrumentation of the dump thread. This would leave the
conditional execution of an internal DEBUG_SYNC command active
(through the usage of DBUG_EXECUTE_IF).
We fix#1 and #2 by waiting for the threads to receive the signal and
only then issue the reset. We fix#3 by reseting the DEBUG variable,
thus deactivating the dump thread internal DEBUG_SYNC command.
The sql_slave_skip_counter is important to be able to recover replication from
certain errors. Often, an appropriate solution is to set
sql_slave_skip_counter to skip over a problem event. But setting
sql_slave_skip_counter produced an error in GTID mode, with a suggestion to
instead set @@gtid_slave_pos to point past the problem event. This however is
not always possible; for example, in case of an INCIDENT event, that event
does not have any GTID to assign to @@gtid_slave_pos.
With this patch, sql_slave_skip_counter now works in GTID mode the same was as
in non-GTID mode. When set, that many initial events are skipped when the SQL
thread starts, plus as many extra events are needed to completely skip any
partially skipped event group. The GTID position is updated to point past the
skipped event(s).
MDEV-6344: mysqldump issues FLUSH TABLES, which gets written into binlog and replicated
Add a --gtid option (for compatibility, the original behaviour is preserved
when --gtid is not used).
With --gtid, --master-data and --dump-slave output the GTID position (the
old-style file/offset position is still output, but commented out). Also, a
CHANGE MASTER TO master_use_gtid=slave_pos is output to ensure a provisioned
slave is configured in GTID, as requested.
Without --gtid, the GTID position is still output, if available, but commented
out.
Also fix MDEV-6344, to avoid FLUSH TABLES getting into the binlog. Otherwise a
mysqldump on a slave server will silently inject a GTID which does not exist
on the master, which is highly undesirable.
Also fix an incorrect error handling around obtaining binlog position with
--master-data (was probably unlikely to trigger in most cases).
Analysis: For some reason table stats for a table pointed from a index
is not initialized. Added additional warning output on this situation
and table stats initialization. This is better than asserting.
These tests use search_pattern_in_file.inc to search the error log for
expected output. However, search_pattern_in_file.inc by default searched only
the first 50000 bytes, so if the error log grew too big the tests would fail.
This patch extends search_pattern_in_file.inc with an option to specify how
much of the file to search, and whether to search from the start of the file
or from the end. Then the rpl.rpl_checksum and rpl.rpl_gtid_errorlog test
cases are fixed to search the last 50000 bytes of the error log, which will
work no matter how large prior tests have made it.
The direct cause of the assertion was missing error handling in
record_gtid(). If ha_commit_trans() fails for the statement commit, there was
missing code to catch the error and do ha_rollback_trans() in this case; this
caused close_thread_tables() to assert.
Normally, this error case is not hit, but in this case it was triggered due to
another bug: When a transaction T1 fails during parallel replication, the code
would signal following transactions that they could start to run without
properly marking the error condition. This caused subsequent transactions to
incorrectly start replicating, only to get an error later during their own
commit step. This was particularly serious if the subsequent transactions were
DDL or MyISAM updates, which cannot be rolled back and would leave replication
in an inconsistent state.
Fixed by 1) in case of error, only signal following transactions to continue
once the error has been properly marked and those transactions will know not
to start; and 2) implement proper error handling in record_gtid() in the case
that statement commit fails.
If replication breaks in GTID mode, it is not trivial to determine the GTID of
the failing event group. This is a problem, as such GTID is needed eg. to
explicitly set @@gtid_slave_pos to skip to after that event group, or to
compare errors on different servers, etc.
Fix by ensuring that relevant slave errors logged to the error log include the
GTID of the event group containing the problem event.
This is MySQL Bug#59123. The message string stored in an INCIDENT event was
not zero-terminated. This caused any following checksum bytes (if enabled on
the master) to be output to the error log as trailing garbage when the message
was printed to the error log.
Backport the patch from MySQL 5.6:
revno: 2876.228.200
revision-id: zhenxing.he@sun.com-20110111051323-w2xnzvcjn46x6h6u
committer: He Zhenxing <zhenxing.he@sun.com>
timestamp: Tue 2011-01-11 13:13:23 +0800
message:
BUG#59123 rpl_stm_binlog_max_cache_size fails sporadically with found warnings
Also add a test case.
Bug#16415173 CRLF INSTEAD OF LF IN SQL-BENCH SCRIPTS
Correct perms and converts from Windows style to UNIX style line endings on some files.
Fix perms on installed ini files.
(MySQL 5.5 version)
- "ANALYZE $stmt" should discard select's output, but it should still
evaluate the output columns (otherwise, subqueries in select list
are not executed)
- SHOW EXPLAIN's code practice of calling JOIN::save_explain_data()
after JOIN::exec() is disastrous for ANALYZE, because it resets
all counters after the first execution. It is stopped
= "Late" test_if_skip_sort_order() calls explicitly update their part
of the query plan.
= Also, I had to rewrite I_S optimization to actually have optimization
and execution stages.
MySQL 5.6 implemented WL#344, which is about a MASTER_DELAY option to CHANGE
MASTER. But as part of this worklog, the format of the realy-log.info file was
changed. The new format is not understood by earlier versions, and nor by
MariaDB 10.0, so changing server to those versions would cause the slave to
abort with an error due to reading incorrect data out of relay-log.info.
Fix this by backporting from the WL#344 patch just the code that understands
the new relay-log.info format. We still write out the old format, and none of
the MASTER_DELAY feature is backported with this commit.
CORRUPTS FRM
Analysis:
---------
ALTER TABLE on a partitioned table resulted in the wrong
engine being written into the table's FRM file and displayed
in SHOW CREATE TABLE.
The prep_alter_part_table() modifies the partition_info object
for TABLE instance representing the old version of table.
If the ALTER TABLE ENGINE statement fails, the partition_info
object for the TABLE contains the altered storage engine name.
The SHOW CREATE TABLE uses the TABLE object to display the table
information, hence displays incorrect storage engine for the table.
Also a subsequent successful ALTER TABLE operation will write the
incorrect engine information into the FRM file.
Fix:
---
A copy of the partition_info object is created before modification so
that any changes would not cause the the original partition_info object
to be modified if the ALTER TABLE fails.(Backported part of the code
provided as fix for bug#14156617 in mysql-5.6.6).
* enum values to index different ACL tables, instead of hard-coded numbers
(even different in diffent functions).
* move TABLE_LIST initialization into open_grant_tables()
and use it everywhere
* change few my_bool's to bool's
* Don't write frm for tmp tables
* pass frm image down to open_table_uncached, when possible
* don't use truncate-by-recreate for temp tables - cannot recreate
without frm, and delete_all_rows is faster anyway
Auto-generate the allowed list of values for enum/set/flagset options
in --help output. But don't do that when the help text already has them.
Also, remove lists of values from help strings of various options, where
they were simply listed without any additional information.
it checked that the default is lz4. Which only worked on systems that
had lz4 and did not have lzo. Now it checks for the default to be zlib,
which works on systems that has neither lz4 or lzo. Like our package
builders in buildbot. This is intentional, we don't want introduce
additional dependencies (lz4, lzo) for our packages just yet.
This can (and will) be reconsidered, and this test can (and will)
be updated again.
The INCIDENT_EVENT always caused slave error and abort, without checking
--slave-skip-errors.
Now, if error 1590, ER_SLAVE_INCIDENT is included in the --slave-skip-errors
list, incident events will be ignored.
This is a merge of this MySQL 5.6 patch:
revision-id: frazer@mysql.com-20110314170916-ypgin17otj3ucx95
committer: Frazer Clement <frazer@mysql.com>
timestamp: Mon 2011-03-14 17:09:16 +0000
message:
Bug#11799671 NOT POSSIBLE TO SKIP INCIDENT ERRORS
DISCONNECT CON1 AND CON2
Problem:
The test suite/binlog/t/binlog_killed.test makes the connections
con1 and con2 but forgets to disconnect them + wait till that
operation is finished at test end.
This mistake has the potential to harm subsequent tests in
case these tests depend on the content of the processlist.
Solution:
Added disconnect + wait_until_disconnected.inc
within the test cleanup.
- Filesort has an optmization where it reads only columns that are
needed before the sorting is done.
- When ref(_or_null) is picked by the join optimizer, it may remove parts
of WHERE clause that are guaranteed to be true.
- However, if we use quick select, we must put all of the range columns into the
read set. Not doing so will may cause us to fail to detect the end of the range.
That particular part of slave connect to master was missing code to handle
retry in case of network errors. The same problem is present in MySQL 5.5, but
fixed in MySQL 5.6.
Fixed with this patch, by adding the code (mostly identical to MySQL 5.6), and
also adding a test case.
I checked other queries done towards master during slave connect, and they now
all seem to handle reconnect in case of network failures.
Adapt default_tmp_storage_engine implementation from mysql-5.6
New feature (as compared to 5.6), default_tmp_storage_engine=NULL
means that temporary tables will use default_storage_engine value.
This makes the behavior backward compatible.
