…info::get_mi

When using JSON fuctions as a parameter to a
SOURCE_POS_WAIT/MASTER_POS_WAIT, something that should also apply to
other functions that return text results, the string extracted from
the parameter did not have a safe pointer to it.

The cause is that the length of the string was not properly marked
inside its allocated space.

Usage of the method c_ptr_safe fixes this issue.

Change-Id: Ic2c54999293aa2e0833594754ad681d7453e03a1

DESCRIPTION:
============
Certain range queries on a table with index prefixed
BLOB/TEXT columns could lead to a server exit.

ANALYSIS:
=========
While opening the table based on its table share, in
open_table_from_share(), we create a copy of the key_info
from TABLE_SHARE object to TABLE object. If the key is
prefixed, we allocate a new Field object, having its
field_length set to the prefix key length, and point the
table's matching key_part->field to this new Field object.
We skip creating the new Field object for prefixed BLOB
columns.

A secondary key is extended by adding primary key parts to
it if the primary key part does not exist in the secondary
key or the key part in the secondary key is a prefix of the
key field (add_pk_parts_to_sk()). The consequence of
skipping the creation of new Field object for prefixed BLOB
columns is that the key parts from the secondary key and
primary key will be pointing to the same Field object.

Later, while performing end-range scan, we check if the key
is within range (compare_key_in_buffer()). We change the
offsets of all the fields in the key range to make the
fields point to the record buffer
(move_key_field_offsets()). In case of BLOBs, we end up
moving the same field twice in move_key_field_offsets().
This leads to accessing out of bound memory while performing
key comparison.

FIX:
====
We allow creating new Field object even for BLOB columns in
open_table_from_share().

Note:
=====
This issue is not a regression but rather was exposed in
5.6.36 by the patch for Bug#23481444: OPTIMISER CALL
ROW_SEARCH_MVCC() AND READ THE INDEX APPLIED BY UNCOMMITTED
ROWS.

Change-Id: I407dec8a997de2c51ebf62351351288beb7dde5e

There are currently two lengths of LQH_TRANSCONF used.

The signal printer should accomodate both, avoiding
printing uninitialised data and not making invalid
assertions about the supported signal lengths.

Change-Id: I0757b512445ae30c310b9c1ec37ed3ae2e1e4e38

Avoid reading signal section pointers that are not
present.

Change-Id: Ifdc5ae688ae2f3d3c64c895ca7a062b1ab0ccc6a

Followup fix - remove [[fallthrough]] not supported on
all compilers used for 7.6.

Change-Id: Ifdc5ae688ae2f3d3c64c895ca7a062b1ab0ccc6a

Originally committed in :
commit 4329a1385304788d642d0dec34174a6cd7842eb7
Author: Frazer Clement <frazer.clement@oracle.com>
Date:   Fri Oct 8 23:54:33 2021 +0100

    Bug #32478380 DEADLOCK TIMEOUT DUE TO PROBLEM IN REDO LOG QUEUE HANDLING

Originally
Approved by : Maitrayi Sabaratnam <maitrayi.sabaratnam@oracle.com>

Change-Id: Idcd159d0da98b6ed4c8542895fcfbaa962a75240

7.6 backport

Improve observability of API failure handling stall

 - QMGR signals blocks yet to complete API failure handling
   to dump block internal API failure handling state
 - TC enhanced to
   - Track + dump API failure handling sub-state
   - Dump info about remaining transactions to be handled
   - Include TC instance number in generated logs
   - Also dump to node log in cases where truncation may
     occur in cluster log

Change-Id: I20c96ba9081610abd4c4f9696bada496b8f4c1ba

…mysql-5.7)

The fprintf_string function present in mysqldump takes the quote
of the string as a parameter, but does not pass it to the
mysql_real_escape_string_quote to escape the string.

The fix is to pass the string quote to mysql_real_escape_string_quote
as a parameter.

Added a test case.

Change-Id: Idc2001a96679fe32bb48e5e3a14d724d5ab9cb9f

Backport to 7.6

Implement a time limit on API failure handling at the data nodes.

QMGR already logs when API failure handling is taking a long time.

This is extended with a time limit on how long a data node is
allowed to take to handle an API node failure before it is
considered a failure of the data node itself.

API failure handling involves cleanup of transaction state
in the DICT, TC, SPJ and SUMA blocks.

DICT transaction state can include schema transactions which may
take some time to complete, so a hard-coded long timeout (7 days)
is used for cases where API failure handling is blocked on DICT
processing.

For other blocks, a shorter configurable timeout is used.

When the timeout elapses, the data node with the problem will
be shutdown, which will hopefully help resolve the problem.

The short timeout is configured with a new data node configuration
parameter :

 ApiFailureHandlingTimeout
   Units : Seconds
   Where :
     0 = No limit
     1..10 = 10 seconds
     > 10 = Limit in seconds
   Default: 600

Two new tests are added to extend coverage of API failure
handling :

  testNodeRestart -n multi_apifail
    Coverage of concurrent API failure handling

  testNodeRestart -n timeout_apifail
    Coverage of data node timeout of api failure handling

Change-Id: Iefea39042cdd4ea83fb22fe2681f9e7bf7d56dba

Backport to 7.6

Two changes :
 1.  Have node error handling set thread watchdog
     state prior to attempting to serialise or log
     error details to files.
     This helps users understand whether Watchdog
     logs indicate a detected overload, or whether
     they indicate a delay in shutting down a
     data node.

 2.  Have the Watchdog thread treat 'slow logging'
     as a special case.
     If a registered thread exceeds its time allowance
     in a shutdown logging state then the watchdog directly
     calls NdbShutdown(), which is more likely to lead
     to an immediate process exit.
     This improves the system's ability to force a
     timely process failure (and subsequent restart)
     potentially at the expense of some logging.

Test coverage by testNodeRestart -n WatchdogSlowShutdown
is enhanced to cover another case.
Error injection coverage of data node shutdown is refactored
to enable future extensions.

Change-Id: I57eabbdb04423409d0aae1b6e548013a7088f4d0

This reverts commit e86c8eaa8012de53d67f4489f6c3774abdc4cdc1.

Change-Id: I1e7d58346e92ef2b36ec9dad5d8f4e6128a72c4b

Change-Id: I446a38597730c6af2d91ea0891be4ad65c099ab3

Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>