- If a subquery is correlated wrt a const table, it will change from
being a "DEPENDENT SUBQUERY" into "SUBQUERY", at the end of its parent's
JOIN::optimize() call. Handle this, update the subquery's QPF.
- Make show_explain.test to work
= "Query plan already deleted" does not happen anymore.
= Handle special case of queries that don't have top-level selects, like
SET x = (SELECT ...)
- Make query plan be re-saved after the first join execution
(saving it after JOIN::cleanup is too late because EXPLAIN output
is currently produced before that)
- Handle QPF allocation/deallocation for edge cases, like unsuccessful
BINLOG command.
- Work around the problem with UNION's direct subselects not being visible.
- Update test results ("Using temporary; Using filesort" are now always printed
last in the Extra column)
- This cset gets rid of memory leaks/crashes. Some result mismatches still remain.
- Update view.result (old EXPLAIN didn't match the execution)
- Put in a stub code to work around the
SELECT ... UNION SELECT ... ORDER BY (subuqery) problem
The test did RESET MASTER and then tried to use --sync_with_master
to wait for GTID-based replication to catch up. This though has
a race, there is a small window where the _old_ pre-RESET MASTER
position on the slave is higher than the new pos-RESET MASTER
position, causing the --sync_with_master to be a no-op.
Fix by using include/wait_condition.inc instead.
When we load the slave state from the mysql.gtid_slave_pos at server start, we
need to load all the rows into the in-memory hash, not just the most recent
one in each replication domain. Otherwise we accumulate cruft in the form of
old rows each time the server restarts.
- Handle another specific case where there the JOIN
never had a query plan, but had multiple join->cleanup(full=true) calls
- The idea that there can only be MAX_TABLES subuqeries/unions was
wrong. Switch QPF_query to using a Dynamic_array.
= make Dynamic_array template support size growth. its underlying
DYNAMIC_ARRAY supports it. (this part will need more polishing)
In record_gtid(), too many rows were deleted from the slave position
hash - we need to always keep on to the most recent committed row,
so we have a valid slave position at all times.
- Handle statements inside SPs:
= regular statements
= SET command, which does not have its own statement.
- Handle execution of subquery from range optimizer:
allocate subquery QPFs on the same MEM_ROOT as the whole
query plan was allocated.
This requires that subselect's footprints are saved before it is deleted.
Attempt to save select's QPF exposes one to a variety of edge cases:
- the select may be a UNION's "fake select" which has no valid id
- optimization may fail in the middle (but subsequent JOIN::optimize() calls
will succeed, despite the fact that there never was a query plan)
Single table UPDATE/DELETE
- Correctly print type=SIMPLE vs type=PRIMARY
- Handle UPDATE/DELETE of mergeable VIEWs: we get the
VIEW's select as the first subquery.
(MySQL 5.6 doesn't print it because it finds that the
subquery is not attached to any select)
- Introduce back QueryPlan/QueryPlanFootprint separation for
single-table UPDATEs/DELETEs
- Create an empty QueryPlanFootprint for all kinds of queries
- Introduce "Query Plan Footprints" (abbrev. QPFs)
QPF is a part of query plan that is
1. sufficient to produce EXPLAIN output,
2. can be used to produce EXPLAIN output even after its subquery/union
was executed and deleted
3. is cheap to save so that we can always save query plans
- This patch doesn't fully address #2, we make/save strings for
a number of EXPLAIN's columns. This will be fixed.
Implement discovery of table non-existence, and related changes:
1. Split GTS_FORCE_DISCOVERY (that was meaning two different things in
two different functions) into GTS_FORCE_DISCOVERY and GTS_USE_DISCOVERY.
2. Move GTS_FORCE_DISCOVERY implementation into open_table_def().
3. In recover_from_failed_open() clear old errors *before* discovery,
not after successful discovery. The final error should come
from the discovery.
4. On forced discovery delete table .frm first. Discovery will write
a new one, if desired.
5. If the frm file exists, but not the table in the engine, force
rediscovery if the engine supports it.
1. default db type for partitions was stored as 1-byte DB_TYPE code,
which doesn't work for dynamically generated codes.
2. storage engine plugin for default db type wasn't locked at all,
which could trivially crash for dynamic plugins.
Now the storage engine name is stored in the extra2 section,
and the plugin is correctly locked.
1. DROP DATABASE should use ha_discover_table_names(), not look at .frm files.
2. filename_to_tablename() also encodes temp file names #sql- -> #mysql50##sql
3. no special treatment for #sql- files, no TABLE_LIST::internal_tmp_table
4. discover also table file names, that start from #
Now whenever we reach the GTID point requested from the slave (when using GTID
position to connect), we send a fake Gtid_list event. This event is used by
the slave to know the current old-style position for MASTER_POS_WAIT(), and
later the similar binlog position for MASTER_GTID_WAIT().
Without this fake event, if the slave is already fully up-to-date with the
master, there may be no events sent at the given position for an indeterminate
time.
If the mysql.gtid_slave_pos table is not available, we cannot load nor update
the current GTID position persistently. This can happen eg. after an upgrade,
before mysql_upgrade_db is run, or if the table is InnoDB and the server is
restarted without the InnoDB storage engine enabled.
Before, replication always failed to start if the table was unavailable. With
this patch, we try to continue with old-style replication, after suitable
complaints in the error log. In strict mode, or if slave is configured to use
GTID, slave still refuses to start.
There was some old code that cleared the position in CHANGE MASTER,
it was forgotten to be removed.
In addition, add code that saves/restores the old-style position
when we nuke the old relay logs as part of GTID slave start.
Normally we will not use these, but it could be useful in case
the GTID connect fails and user wants to go back to the old-style
coordinates.
mysql-test/r/servers.result:
Added test case for MDEV-4594.
mysql-test/t/servers.test:
Added test case for MDEV-4594.
sql/mysqld.cc:
Move servers_init() call to init_server_components(), so it is called
for embedded as well.
Call servers_free() even if NO_EMBEDDED_ACCESS_CHECKS defined (because
we call servers_init() anyway).