Where a read-only server permits writes through replication, it
should not permit user connections to commit/rollback XA
transactions prepared via replication. The bug reported in
MDEV-30978 shows that this can happen. This is because there is no
read only check in the XA transaction logic, the most relevant one
occurs in ha_commit_trans() for normal statements/transactions.
This patch extends the XA transaction logic to check the read only
status of the server before performing an XA COMMIT or ROLLBACK.
Reviewed By:
Andrei Elkin <andrei.elkin@mariadb.com>
rw_trx_hash_t::find() acquires element->mutex, then unpins pins, used for
lf_hash element search. After that the "element" can be deallocated and
reused by some other thread.
If we take a look rw_trx_hash_t::insert()->lf_hash_insert()->lf_alloc_new()
calls, we will not find any element->mutex acquisition, as it was not
initialized yet before it's allocation. rw_trx_hash_t::insert() can reuse
the chunk, unpinned in rw_trx_hash_t::find().
The scenario is the following:
1. Thread 1 have just executed lf_hash_search() in
rw_trx_hash_t::find(), but have not acquired element->mutex yet.
2. Thread 2 have removed the element from hash table with
rw_trx_hash_t::erase() call.
3. Thread 1 acquired element->mutex and unpinned pin 2 pin with
lf_hash_search_unpin(pins) call.
4. Some thread purged memory of the element.
5. Thread 3 reused the memory for the element, filled element->id,
element->trx.
6. Thread 1 crashes with failed "DBUG_ASSERT(trx_id == trx->id)"
assertion.
Note that trx_t objects are also reused, see the code around trx_pools
for details.
The fix is to invoke "lf_hash_search_unpin(pins);" after element->trx is
stored in local variable in rw_trx_hash_t::find().
Reviewed by: Nikita Malyavin, Marko Mäkelä.
The user XA commit execution branch was caught not have been covered
with MDEV-21953 fixes.
The XA involved deadlock is resolved now to apply the former fixes
pattern.
Along the fixes the following changes have been implemented.
- MDL lock attribute correction
- dissociation of the externally completed XA from the current
thread's xid_state in the error branches
- cleanup_context() preseves the prepared XA
- wait_for_prior_commit() is relocated to satisfy both
the binlog ON (log-slave-updates and skip-log-bin)
and OFF slave execution branches.
The real fix for MDEV-15532 will be pushed into 10.2 and 10.6
This is an additional fix for 10.4.
In 10.4 trans_xa_detach was introduced. However THD::cleanup() assumes
that after trans_xa_detach() is done, there is no registered transactions
anymore. In the 10.2 patch there will be an assert to ensure this, which
will cause 10.4 to fail.
The fix used is to reset the transaction flags in trans_xa_detach().
MDEV-21953 deadlock between BACKUP STAGE BLOCK_COMMIT and parallel
replication
Fixed by partly reverting MDEV-21953 to put back MDL_BACKUP_COMMIT locking
before log_and_order.
The original problem for MDEV-21953 was that while a thread was waiting in
for another threads to commit in 'log_and_order', it had the
MDL_BACKUP_COMMIT lock. The backup thread was waiting to get the
MDL_BACKUP_WAIT_COMMIT lock, which blocks all new MDL_BACKUP_COMMIT locks.
This causes a deadlock as the waited-for thread can never get past the
MDL_BACKUP_COMMIT lock in ha_commit_trans.
The main part of the bug fix is to release the MDL_BACKUP_COMMIT lock while
a thread is waiting for other 'previous' threads to commit. This ensures
that no transactional thread keeps MDL_BACKUP_COMMIT while waiting, which
ensures that there are no deadlocks anymore.
This happend when using XA transactions. I also added some extra asserts
to ensure that m_transactions are properly cleared.
Other things:
- Removed set_time() from THD::init_for_queries() as dispatch_command()
is already doing that.
- Removed duplicate init_for_queries() from prepare_new_connection_state().
The init_for_queries() functions should only be called once per
connection.
The issue was:
T1, a parallel slave worker thread, is waiting for another worker thread to
commit. While waiting, it has the MDL_BACKUP_COMMIT lock.
T2, working for mariabackup, is doing BACKUP STAGE BLOCK_COMMIT and blocks
all commits.
This causes a deadlock as the thread T1 is waiting for can't commit.
Fixed by moving locking of MDL_BACKUP_COMMIT from ha_commit_trans() to
commit_one_phase_2()
Other things:
- Added a new argument to ha_comit_one_phase() to signal if the
transaction was a write transaction.
- Ensured that ha_maria::implicit_commit() is always called under
MDL_BACKUP_COMMIT. This code is not needed in 10.5
- Ensure that MDL_Request values 'type' and 'ticket' are always
initialized. This makes it easier to check the state of the MDL_Request.
- Moved thd->store_globals() earlier in handle_rpl_parallel_thread() as
thd->init_for_queries() could use a MDL that could crash if store_globals
where not called.
- Don't call ha_enable_transactions() in THD::init_for_queries() as this
is both slow (uses MDL locks) and not needed.
All changes (except one) is of type
thd->transaction. -> thd->transaction->
thd->transaction points by default to 'thd->default_transaction'
This allows us to 'easily' have multiple active transactions for a
THD object, like when reading data from the mysql.proc table
Lifted long standing limitation to the XA of rolling it back at the
transaction's
connection close even if the XA is prepared.
Prepared XA-transaction is made to sustain connection close or server
restart.
The patch consists of
- binary logging extension to write prepared XA part of
transaction signified with
its XID in a new XA_prepare_log_event. The concusion part -
with Commit or Rollback decision - is logged separately as
Query_log_event.
That is in the binlog the XA consists of two separate group of
events.
That makes the whole XA possibly interweaving in binlog with
other XA:s or regular transaction but with no harm to
replication and data consistency.
Gtid_log_event receives two more flags to identify which of the
two XA phases of the transaction it represents. With either flag
set also XID info is added to the event.
When binlog is ON on the server XID::formatID is
constrained to 4 bytes.
- engines are made aware of the server policy to keep up user
prepared XA:s so they (Innodb, rocksdb) don't roll them back
anymore at their disconnect methods.
- slave applier is refined to cope with two phase logged XA:s
including parallel modes of execution.
This patch does not address crash-safe logging of the new events which
is being addressed by MDEV-21469.
CORNER CASES: read-only, pure myisam, binlog-*, @@skip_log_bin, etc
Are addressed along the following policies.
1. The read-only at reconnect marks XID to fail for future
completion with ER_XA_RBROLLBACK.
2. binlog-* filtered XA when it changes engine data is regarded as
loggable even when nothing got cached for binlog. An empty
XA-prepare group is recorded. Consequent Commit-or-Rollback
succeeds in the Engine(s) as well as recorded into binlog.
3. The same applies to the non-transactional engine XA.
4. @@skip_log_bin=OFF does not record anything at XA-prepare
(obviously), but the completion event is recorded into binlog to
admit inconsistency with slave.
The following actions are taken by the patch.
At XA-prepare:
when empty binlog cache - don't do anything to binlog if RO,
otherwise write empty XA_prepare (assert(binlog-filter case)).
At Disconnect:
when Prepared && RO (=> no binlogging was done)
set Xid_cache_element::error := ER_XA_RBROLLBACK
*keep* XID in the cache, and rollback the transaction.
At XA-"complete":
Discover the error, if any don't binlog the "complete",
return the error to the user.
Kudos
-----
Alexey Botchkov took to drive this work initially.
Sergei Golubchik, Sergei Petrunja, Marko Mäkelä provided a number of
good recommendations.
Sergei Voitovich made a magnificent review and improvements to the code.
They all deserve a bunch of thanks for making this work done!
MDEV-21854 xa commit `xid` one phase for already prepared transaction must always error out
Added state and one-phase option checks to XA "external" commit/rollback
branches. While the XA standard does not prohibit it,
Commit and Rollback of an XA external to the current ongoing transaction
is not allowed; after all the current transaction may rollback
to not being able to revert that decision.
XA specification doesn't permit empty gtrid. It is now enforced by this
patch. This solution was agreed in favour of fixing InnoDB, which doesn't
expect empty XID since early 10.5.
Also fixed wrong assertion (and added a test cases) that didn't permit
64 bytes gtrid + 64 bytes bqual.
Let xid_cache_insert()/xid_cache_delete() handle xa_state.
Let session tracker use is_explicit_XA() rather than xa_state != XA_NOTR.
Fixed open_tables() to refuse data access in XA_ROLLBACK_ONLY state.
Removed dead code from THD::cleanup(). It was supposed to be a reminder,
but it got messed up over time.
spider_internal_start_trx() is called either with XA_NOTR or XA_ACTIVE,
which is guarded by server callers. Thus is_explicit_XA() is acceptable
replacement for XA_ACTIVE check (which was likely wrong anyway).
Setting xa_state to XA_PREPARED in spider_internal_xa_prepare() isn't
meaningful, as this value is never accessed later. It can't be accessed
by current thread and it can't be recovered either. It can only be
accessed by spider internally, which never happens.
Make spider_xa_lock()/spider_xa_unlock() static.
Part of MDEV-7974 - backport fix for mysql bug#12161 (XA and binlog)
XID_STATE::rm_error is never used by internal 2PC, it is intended to be
used by explicit XA only.
Also removed redundant xid reset from THD::init_for_queries(). Must've
been done already either by THD::transaction constructor or by
THD::cleanup().
Part of MDEV-7974 - backport fix for mysql bug#12161 (XA and binlog)
