The COM_STMT_FETCH command will create a response. This was a
readwritesplit-specific interpretation of the command and it was wrong.
Also record the currently executed command event for session commands.
Readwritesplit would not handle multiple overlapping COM_STMT_EXECUTE
commands properly if they opened cursors. This was due to the fact that
the result would not be marked as complete and COM_STMT_FETCH commands
were executed as if they did not return results.
The correct implementation is to consider a COM_STMT_EXECUTE that opens a
cursor complete only when the first EOF packet is read (that is, when the
resultset header is read). This allows subsequent COM_STMT_FETCH commands
to be handled separately.
The separate COM_STMT_FETCH handling must count the number of packets that
are being fetched. This allows correct tracking of the state of a
COM_STMT_FETCH by checking that the number of packets is correct or the
second EOF/ERR packet is read.
When a LOAD DATA LOCAL INFILE is actively rejected by the server, the
server sends an error to the client. This error was not detected and the
router was stuck in the special mode that handles LOAD DATA LOCAL INFILE.
The use of `router_options=master,slave` was not working as expected. This
was mostly caused by the master bit checks using a bitwise AND instead of
comparing equality. In addition to this, the master would not be
considered a valid candidate if both slaves and masters were available.
When a multi-statement query consisting completely of UPDATE statements is
received, the packets can be received in two separate buffers. To cope
with this situation, the state change into REPLY_STATE_RSET_COLDEF must
only be done if the buffer contains more than a single packet.
When readwritesplit receives a reply from a backend, an info level log
message is now logged. This allows easier debugging of situations where
replies aren't properly returned by the router.
If the avrorouter is being build and the required libraries are not found,
the configuration process should fail. Adding the command to bypass this
into the error message should make it easier to disable this part if it is
not needed.
After a temporary table is created, readwritesplit will check whether a
query drops or targets that temporary table. The check for query type was
missing from the table dropping part of the code. The temporary table read
part was checking that the query is a text form query.
Added a debug assertion to the query parsing function in qc_sqlite to
catch this type of interface misuse.
If a server is removed from a service, readconnroute will not verify that
the server it is connected to is still the same root master. This fixes
the regression of MXS-1418.
When a BEGIN statement is prepared using the binary protocol, it returns a
single OK packet. Due to a bug in the code that deals with multi-statement
results and EOF packets, the response was never sent to the client.
Also added back the error messages of failed session commands to the INFO
level. This way it's still possible to see why a session command fails but
the log isn't flooded by them in normal usage.
The responses of slaves that arrived before the master were always
compared to the empty value of 0x00. If the slave connection replied after
the master, the comparison was correct.
This commit introduces a map of slaves and their responses that
are handled once the master's response arrives.
Removed false error message about failed session commands. An error in
response to a session command is a perfectly valid result.
Also added the explicit commands that the master and slave return to the
warning that is logged when the results differ.
The debug assertion wasn't well placed as it is perfectly possible that a
master connnection exists but it is not in use. This can be further
checked by asserting that the master is indeed closed and not in use.
Moved the original debug assertion into a separate branch that should
catch any errors in the routing logic.
The token skipping function did not check for a period or an opening
parenthesis when parsing the test. Also fixed a debug assertion when only
NULL values were inserted.
As chages to the transaction state are detected by the protocol level
mini-parser, there's no need to fully classify queries inside read-only
transactions. This should be a good performance boost for loads that
heavily use read-only transactions.
Used the correct value in table_create_alloc and remove unused
parameter. Use the pre-calculated end pointer when looking for events.
Always use the column count of the TABLE_MAP event as all mismatches are
detected earlier.
The parser checks whether the FIRST or AFTER keywords are used and, if
AFTER is used, extracts the relevant column name.
Added a test case that checks that the parsing works and detects the
correct column names.
According to customer reports collecting the statistics has a significant
impact on the performance. As we don't need that information we can just
as well turn off that.
Further, since maxscale-common now links to the sqlite3-library, no
module needs to do that explicitly.
The sprintf calls failed due to a warning about possible buffer
overflow. Curiously enough, the same warnings do appear on Fedora 26 but
only when the calls are changed to snprintf.
The warning that a schema already exists is obsolete as mapped tables are
now always opened instead of being reused. This causes the schema checks
to be done for each mapped table.
When an assertion fails due to an overflow of the event buffer, all
processed values for that event are dumped.
This commit also enables the assertions even for non-debug builds which
should speed up the elimination process for bugs in the avrorouter. The
overhead of doing this is minimal as the output is already gathered for
the INFO level logging.
Make all modules lowercase and make module loading case
insensitive. Further, make command invocation case insensitive,
as far as the module name is conserned.
The `purge` command can be used to reset the conversion
process. Currently, executing the `purge` module command and restarting
MaxScale is the only correct way to reset the conversion process.
Added more debug assertions to JSON serialization of the Avro schema. Also
checked that the column count for the TABLE_CREATE object matches that of
the TABLE_MAP object.
