MDEV-11581: Mariadb starts InnoDB encryption threads
when key has not changed or data scrubbing turned off
Background: Key rotation is based on background threads
(innodb-encryption-threads) periodically going through
all tablespaces on fil_system. For each tablespace
current used key version is compared to max key age
(innodb-encryption-rotate-key-age). This process
naturally takes CPU. Similarly, in same time need for
scrubbing is investigated. Currently, key rotation
is fully supported on Amazon AWS key management plugin
only but InnoDB does not have knowledge what key
management plugin is used.
This patch re-purposes innodb-encryption-rotate-key-age=0
to disable key rotation and background data scrubbing.
All new tables are added to special list for key rotation
and key rotation is based on sending a event to
background encryption threads instead of using periodic
checking (i.e. timeout).
fil0fil.cc: Added functions fil_space_acquire_low()
to acquire a tablespace when it could be dropped concurrently.
This function is used from fil_space_acquire() or
fil_space_acquire_silent() that will not print
any messages if we try to acquire space that does not exist.
fil_space_release() to release a acquired tablespace.
fil_space_next() to iterate tablespaces in fil_system
using fil_space_acquire() and fil_space_release().
Similarly, fil_space_keyrotation_next() to iterate new
list fil_system->rotation_list where new tables.
are added if key rotation is disabled.
Removed unnecessary functions fil_get_first_space_safe()
fil_get_next_space_safe()
fil_node_open_file(): After page 0 is read read also
crypt_info if it is not yet read.
btr_scrub_lock_dict_func()
buf_page_check_corrupt()
buf_page_encrypt_before_write()
buf_merge_or_delete_for_page()
lock_print_info_all_transactions()
row_fts_psort_info_init()
row_truncate_table_for_mysql()
row_drop_table_for_mysql()
Use fil_space_acquire()/release() to access fil_space_t.
buf_page_decrypt_after_read():
Use fil_space_get_crypt_data() because at this point
we might not yet have read page 0.
fil0crypt.cc/fil0fil.h: Lot of changes. Pass fil_space_t* directly
to functions needing it and store fil_space_t* to rotation state.
Use fil_space_acquire()/release() when iterating tablespaces
and removed unnecessary is_closing from fil_crypt_t. Use
fil_space_t::is_stopping() to detect when access to
tablespace should be stopped. Removed unnecessary
fil_space_get_crypt_data().
fil_space_create(): Inform key rotation that there could
be something to do if key rotation is disabled and new
table with encryption enabled is created.
Remove unnecessary functions fil_get_first_space_safe()
and fil_get_next_space_safe(). fil_space_acquire()
and fil_space_release() are used instead. Moved
fil_space_get_crypt_data() and fil_space_set_crypt_data()
to fil0crypt.cc.
fsp_header_init(): Acquire fil_space_t*, write crypt_data
and release space.
check_table_options()
Renamed FIL_SPACE_ENCRYPTION_* TO FIL_ENCRYPTION_*
i_s.cc: Added ROTATING_OR_FLUSHING field to
information_schema.innodb_tablespace_encryption
to show current status of key rotation.
In the 10.1 InnoDB Plugin, a call os_event_free(buf_flush_event) was
misplaced. The event could be signalled by rollback of resurrected
transactions while shutdown was in progress. This bug was caught
by cmake -DWITH_ASAN testing. This call was only present in the
10.1 InnoDB Plugin, not in other versions, or in XtraDB.
That said, the bug affects all InnoDB versions. Shutdown assumes the
cessation of any page-dirtying activity, including the activity of
the background rollback thread. InnoDB only waited for the background
rollback to finish as part of a slow shutdown (innodb_fast_shutdown=0).
The default is a clean shutdown (innodb_fast_shutdown=1). In a scenario
where InnoDB is killed, restarted, and shut down soon enough, the data
files could become corrupted.
logs_empty_and_mark_files_at_shutdown(): Wait for the
rollback to finish, except if innodb_fast_shutdown=2
(crash-like shutdown) was requested.
trx_rollback_or_clean_recovered(): Before choosing the next
recovered transaction to roll back, terminate early if non-slow
shutdown was initiated. Roll back everything on slow shutdown
(innodb_fast_shutdown=0).
srv_innodb_monitor_mutex: Declare as static, because the mutex
is only used within one module.
After each call to os_event_free(), ensure that the freed event
is not reachable via global variables, by setting the relevant
variables to NULL.
Also, implement MDEV-11027 a little differently from 5.5 and 10.0:
recv_apply_hashed_log_recs(): Change the return type back to void
(DB_SUCCESS was always returned).
Report progress also via systemd using sd_notifyf().
Also, implement MDEV-11027 a little differently from 5.5:
recv_sys_t::report(ib_time_t): Determine whether progress should
be reported.
recv_apply_hashed_log_recs(): Rename the parameter to last_batch.
If page_compression (introduced in MariaDB Server 10.1) is enabled,
the logical action is to not preallocate space to the data files,
but to only logically extend the files with zeroes.
fil_create_new_single_table_tablespace(): Create smaller files for
ROW_FORMAT=COMPRESSED tables, but adhere to the minimum file size of
4*innodb_page_size.
fil_space_extend_must_retry(), os_file_set_size(): On Windows,
use SetFileInformationByHandle() and FILE_END_OF_FILE_INFO,
which depends on bumping _WIN32_WINNT to 0x0600.
FIXME: The files are not yet set up as sparse, so
this will currently end up physically extending (preallocating)
the files, wasting storage for unused pages.
os_file_set_size(): Add the parameter "bool sparse=false" to declare
that the file is to be extended logically, instead of being preallocated.
The only caller with sparse=true is
fil_create_new_single_table_tablespace().
(The system tablespace cannot be created with page_compression.)
fil_space_extend_must_retry(), os_file_set_size(): Outside Windows,
use ftruncate() to extend files that are supposed to be sparse.
On systems where ftruncate() is limited to files less than 4GiB
(if there are any), fil_space_extend_must_retry() retains the
old logic of physically extending the file.
The function trx_purge_stop() was calling os_event_reset(purge_sys->event)
before calling rw_lock_x_lock(&purge_sys->latch). The os_event_set()
call in srv_purge_coordinator_suspend() is protected by that X-latch.
It would seem a good idea to consistently protect both os_event_set()
and os_event_reset() calls with a common mutex or rw-lock in those
cases where os_event_set() and os_event_reset() are used
like condition variables, tied to changes of shared state.
For each os_event_t, we try to document the mutex or rw-lock that is
being used. For some events, frequent calls to os_event_set() seem to
try to avoid hangs. Some events are never waited for infinitely, only
timed waits, and os_event_set() is used for early termination of these
waits.
os_aio_simulated_put_read_threads_to_sleep(): Define as a null macro
on other systems than Windows. TODO: remove this altogether and disable
innodb_use_native_aio on Windows.
os_aio_segment_wait_events[]: Initialize only if innodb_use_native_aio=0.
Remove the debug parameter innodb_force_recovery_crash that was
introduced into MySQL 5.6 by me in WL#6494 which allowed InnoDB
to resize the redo log on startup.
Let innodb.log_file_size actually start up the server, but ensure
that the InnoDB storage engine refuses to start up in each of the
scenarios.
srv_release_threads(): Actually wait for the threads to resume
from suspension. On CentOS 5 and possibly other platforms,
os_event_set() may be lost.
srv_resume_thread(): A counterpart of srv_suspend_thread().
Optionally wait for the event to be set, optionally with a timeout,
and then release the thread from suspension.
srv_free_slot(): Unconditionally suspend the thread. It is always
in resumed state when this function is entered.
srv_active_wake_master_thread_low(): Only call os_event_set().
srv_purge_coordinator_suspend(): Use srv_resume_thread() instead
of the complicated logic.
InnoDB would refuse to start up if there is a mismatch on
the size of the system tablespace files. However, before this
check is conducted, the system tablespace may already have been
heavily modified.
InnoDB should perform the size check as early as possible.
recv_recovery_from_checkpoint_finish():
Move the recv_apply_hashed_log_recs() call to
innobase_start_or_create_for_mysql().
innobase_start_or_create_for_mysql(): Test the mutex functionality
before doing anything else. Use a compile_time_assert() for a
sizeof() constraint. Check the size of the system tablespace as
early as possible.
MariaDB 10.0/MySQL 5.6 using innodb-page-size!=16K
The storage format of FSP_SPACE_FLAGS was accidentally broken
already in MariaDB 10.1.0. This fix is bringing the format in
line with other MySQL and MariaDB release series.
Please refer to the comments that were added to fsp0fsp.h
for details.
This is an INCOMPATIBLE CHANGE that affects users of
page_compression and non-default innodb_page_size. Upgrading
to this release will correct the flags in the data files.
If you want to downgrade to earlier MariaDB 10.1.x, please refer
to the test innodb.101_compatibility how to reset the
FSP_SPACE_FLAGS in the files.
NOTE: MariaDB 10.1.0 to 10.1.20 can misinterpret
uncompressed data files with innodb_page_size=4k or 64k as
compressed innodb_page_size=16k files, and then probably fail
when trying to access the pages. See the comments in the
function fsp_flags_convert_from_101() for detailed analysis.
Move PAGE_COMPRESSION to FSP_SPACE_FLAGS bit position 16.
In this way, compressed innodb_page_size=16k tablespaces will not
be mistaken for uncompressed ones by MariaDB 10.1.0 to 10.1.20.
Derive PAGE_COMPRESSION_LEVEL, ATOMIC_WRITES and DATA_DIR from the
dict_table_t::flags when the table is available, in
fil_space_for_table_exists_in_mem() or fil_open_single_table_tablespace().
During crash recovery, fil_load_single_table_tablespace() will use
innodb_compression_level for the PAGE_COMPRESSION_LEVEL.
FSP_FLAGS_MEM_MASK: A bitmap of the memory-only fil_space_t::flags
that are not to be written to FSP_SPACE_FLAGS. Currently, these will
include PAGE_COMPRESSION_LEVEL, ATOMIC_WRITES and DATA_DIR.
Introduce the macro FSP_FLAGS_PAGE_SSIZE(). We only support
one innodb_page_size for the whole instance.
When creating a dummy tablespace for the redo log, use
fil_space_t::flags=0. The flags are never written to the redo log files.
Remove many FSP_FLAGS_SET_ macros.
dict_tf_verify_flags(): Remove. This is basically only duplicating
the logic of dict_tf_to_fsp_flags(), used in a debug assertion.
fil_space_t::mark: Remove. This flag was not used for anything.
fil_space_for_table_exists_in_mem(): Remove the unnecessary parameter
mark_space, and add a parameter for table flags. Check that
fil_space_t::flags match the table flags, and adjust the (memory-only)
flags based on the table flags.
fil_node_open_file(): Remove some redundant or unreachable conditions,
do not use stderr for output, and avoid unnecessary server aborts.
fil_user_tablespace_restore_page(): Convert the flags, so that the
correct page_size will be used when restoring a page from the
doublewrite buffer.
fil_space_get_page_compressed(), fsp_flags_is_page_compressed(): Remove.
It suffices to have fil_space_is_page_compressed().
FSP_FLAGS_WIDTH_DATA_DIR, FSP_FLAGS_WIDTH_PAGE_COMPRESSION_LEVEL,
FSP_FLAGS_WIDTH_ATOMIC_WRITES: Remove, because these flags do not
exist in the FSP_SPACE_FLAGS but only in memory.
fsp_flags_try_adjust(): New function, to adjust the FSP_SPACE_FLAGS
in page 0. Called by fil_open_single_table_tablespace(),
fil_space_for_table_exists_in_mem(), innobase_start_or_create_for_mysql()
except if --innodb-read-only is active.
fsp_flags_is_valid(ulint): Reimplement from the scratch, with
accurate comments. Do not display any details of detected
inconsistencies, because the output could be confusing when
dealing with MariaDB 10.1.x data files.
fsp_flags_convert_from_101(ulint): Convert flags from buggy
MariaDB 10.1.x format, or return ULINT_UNDEFINED if the flags
cannot be in MariaDB 10.1.x format.
fsp_flags_match(): Check the flags when probing files.
Implemented based on fsp_flags_is_valid()
and fsp_flags_convert_from_101().
dict_check_tablespaces_and_store_max_id(): Do not access the
page after committing the mini-transaction.
IMPORT TABLESPACE fixes:
AbstractCallback::init(): Convert the flags.
FetchIndexRootPages::operator(): Check that the tablespace flags match the
table flags. Do not attempt to convert tablespace flags to table flags,
because the conversion would necessarily be lossy.
PageConverter::update_header(): Write back the correct flags.
This takes care of the flags in IMPORT TABLESPACE.
Sometimes innodb_data_file_size_debug was reported as INT UNSIGNED
instead of BIGINT UNSIGNED. Make it uint instead of ulong to get
a more deterministic result.
InnoDB shutdown failed to properly take fil_crypt_thread() into account.
The encryption threads were signalled to shut down together with other
non-critical tasks. This could be much too early in case of slow shutdown,
which could need minutes to complete the purge. Furthermore, InnoDB
failed to wait for the fil_crypt_thread() to actually exit before
proceeding to the final steps of shutdown, causing the race conditions.
Furthermore, the log_scrub_thread() was shut down way too early.
Also it should remain until the SRV_SHUTDOWN_FLUSH_PHASE.
fil_crypt_threads_end(): Remove. This would cause the threads to
be terminated way too early.
srv_buf_dump_thread_active, srv_dict_stats_thread_active,
lock_sys->timeout_thread_active, log_scrub_thread_active,
srv_monitor_active, srv_error_monitor_active: Remove a race condition
between startup and shutdown, by setting these in the startup thread
that creates threads, not in each created thread. In this way, once the
flag is cleared, it will remain cleared during shutdown.
srv_n_fil_crypt_threads_started, fil_crypt_threads_event: Declare in
global rather than static scope.
log_scrub_event, srv_log_scrub_thread_active, log_scrub_thread():
Declare in static rather than global scope. Let these be created by
log_init() and freed by log_shutdown().
rotate_thread_t::should_shutdown(): Do not shut down before the
SRV_SHUTDOWN_FLUSH_PHASE.
srv_any_background_threads_are_active(): Remove. These checks now
exist in logs_empty_and_mark_files_at_shutdown().
logs_empty_and_mark_files_at_shutdown(): Shut down the threads in
the proper order. Keep fil_crypt_thread() and log_scrub_thread() alive
until SRV_SHUTDOWN_FLUSH_PHASE, and check that they actually terminate.
fil_space_t::recv_size: New member: recovered tablespace size in pages;
0 if no size change was read from the redo log,
or if the size change was implemented.
fil_space_set_recv_size(): New function for setting space->recv_size.
innodb_data_file_size_debug: A debug parameter for setting the system
tablespace size in recovery even when the redo log does not contain
any size changes. It is hard to write a small test case that would
cause the system tablespace to be extended at the critical moment.
recv_parse_log_rec(): Note those tablespaces whose size is being changed
by the redo log, by invoking fil_space_set_recv_size().
innobase_init(): Correct an error message, and do not require a larger
innodb_buffer_pool_size when starting up with a smaller innodb_page_size.
innobase_start_or_create_for_mysql(): Allow startup with any initial
size of the ibdata1 file if the autoextend attribute is set. Require
the minimum size of fixed-size system tablespaces to be 640 pages,
not 10 megabytes. Implement innodb_data_file_size_debug.
open_or_create_data_files(): Round the system tablespace size down
to pages, not to full megabytes, (Our test truncates the system
tablespace to more than 800 pages with innodb_page_size=4k.
InnoDB should not imagine that it was truncated to 768 pages
and then overwrite good pages in the tablespace.)
fil_flush_low(): Refactored from fil_flush().
fil_space_extend_must_retry(): Refactored from
fil_extend_space_to_desired_size().
fil_mutex_enter_and_prepare_for_io(): Extend the tablespace if
fil_space_set_recv_size() was called.
The test case has been successfully run with all the
innodb_page_size values 4k, 8k, 16k, 32k, 64k.
Replace all exit() calls in InnoDB with abort() [possibly via ut_a()].
Calling exit() in a multi-threaded program is problematic also for
the reason that other threads could see corrupted data structures
while some data structures are being cleaned up by atexit() handlers
or similar.
In the long term, all these calls should be replaced with something
that returns an error all the way up the call stack.
Make some global fil_crypt_ variables static.
fil_close(): Call mutex_free(&fil_system->mutex) also in InnoDB, not
only in XtraDB. In InnoDB, sync_close() was called before fil_close().
innobase_shutdown_for_mysql(): Call fil_close() before sync_close(),
similar to XtraDB shutdown.
fil_space_crypt_cleanup(): Call mutex_free() to pair with
fil_space_crypt_init().
fil_crypt_threads_cleanup(): Call mutex_free() to pair with
fil_crypt_threads_init().
Reduce the number of calls to encryption_get_key_get_latest_version
when doing key rotation with two different methods:
(1) We need to fetch key information when tablespace not yet
have a encryption information, invalid keys are handled now
differently (see below). There was extra call to detect
if key_id is not found on key rotation.
(2) If key_id is not found from encryption plugin, do not
try fetching new key_version for it as it will fail anyway.
We store return value from encryption_get_key_get_latest_version
call and if it returns ENCRYPTION_KEY_VERSION_INVALID there
is no need to call it again.
Analysis: By design InnoDB was reading first page of every .ibd file
at startup to find out is tablespace encrypted or not. This is
because tablespace could have been encrypted always, not
encrypted newer or encrypted based on configuration and this
information can be find realible only from first page of .ibd file.
Fix: Do not read first page of every .ibd file at startup. Instead
whenever tablespace is first time accedded we will read the first
page to find necessary information about tablespace encryption
status.
TODO: Add support for SYS_TABLEOPTIONS where all table options
encryption information included will be stored.
we can see from the hang stacktrace, srv_monitor_thread is blocked
when getting log_sys::mutex, so that sync_arr_wake_threads_if_sema_free
cannot get a change to break the mutex deadlock.
The fix is simply removing any mutex wait in srv_monitor_thread.
Patch is reviewed by Sunny over IM.
Backport pull request #125 from grooverdan/MDEV-8923_innodb_buffer_pool_dump_pct to 10.0
WL#6504 InnoDB buffer pool dump/load enchantments
This patch consists of two parts:
1. Dump only the hottest N% of the buffer pool(s)
2. Prevent hogging the server duing BP load
From MySQL - commit b409342c43ce2edb68807100a77001367c7e6b8e
Add testcases for innodb_buffer_pool_dump_pct_basic.
Part of the code authored by Daniel Black
- Make accelerated checksum available to InnoDB and XtraDB.
- Fall back to slice-by-eight if not available. The mode used is printed on startup.
- Will only build on POWER systems at the moment until CMakeLists are modified
to only add the crc32_power8/ files when building on POWER.
running MySQL-5.7 unittest/gunit/innodb/ut0crc32-t
Before:
1..2
Using software crc32 implementation, CPU is little-endian
ok 1
Using software crc32 implementation, CPU is little-endian
normal CRC32: real 0.148006 sec
normal CRC32: user 0.148000 sec
normal CRC32: sys 0.000000 sec
big endian CRC32: real 0.144293 sec
big endian CRC32: user 0.144000 sec
big endian CRC32: sys 0.000000 sec
ok 2
After:
1..2
Using POWER8 crc32 implementation, CPU is little-endian
ok 1
Using POWER8 crc32 implementation, CPU is little-endian
normal CRC32: real 0.008097 sec
normal CRC32: user 0.008000 sec
normal CRC32: sys 0.000000 sec
big endian CRC32: real 0.147043 sec
big endian CRC32: user 0.144000 sec
big endian CRC32: sys 0.000000 sec
ok 2
Author CRC32 ASM code: Anton Blanchard <anton@au.ibm.com>
ref: https://github.com/antonblanchard/crc32-vpmsum
Signed-off-by: Daniel Black <daniel.black@au.ibm.com>
WL#6504 InnoDB buffer pool dump/load enchantments
This patch consists of two parts:
1. Dump only the hottest N% of the buffer pool(s)
2. Prevent hogging the server duing BP load
From MySQL - commit b409342c43ce2edb68807100a77001367c7e6b8e
Analysis: Current implementation will write and read at least one block
(sort_buffer_size bytes) from disk / index even if that block does not
contain any records.
Fix: Avoid writing / reading empty blocks to temporary files (disk).
