The monitor queried the session-specific domain id, which does not follow the global
value while the session is alive. This caused the monitor to follow the wrong gtid
domain if the domain was changed after MaxScale was started. This patch modifies the
query to read the global value instead. Even this is not fool-proof, as existing
sessions can issue writes with the old domain, confusing the gtid-parsing.
Provides a clearer separation between what deals with query
classification and what deals with query routing.
Functions have only been moved. No other cleanup has been
done.
With the configuration entry
dump_last_statements=[never|on_close|on_error]
you can now specify when and if to dump the last statements
of of a session.
With the configuration entry
retain_last_statements=<unsigned>
or the debug flag '--debug=retain-last-statements=<unsigned>',
MaxScale will store the specified number of last statements
for each session. By calling
session_dump_statements(session);
MaxScale will dump the last statements as NOTICE messages.
For debugging purposes.
- session_set_response() made const correct
- set_response() function added to mxs::FilterSession; calls
session_set_response().
- Cache uses set_response() for delivering the cache result
to the client.
We need to copy some data from a AF_UNIX based listener dcb
to the accepted client dcb, to prevent assertion violation in
dcb_get_port(). Further, to be able to log the path in the case
of an authentication error we need to copy that as well.
If a table/database rule has been provided then if the resultset
does not contain table/database names, then we consider it a match
(subject to the column obviously).
Otherwise a rule like
{
"replace": {
"table": "info",
"column": "email"
},
"with": {
"fill": "*"
}
}
could be bypassed with a statement like
SELECT * FROM info UNION SELECT * from info
as the resultset in that case will not indicate that the column emain
is from info, which it will if the statement is
SELECT * FROM info;
Renamed to MariaDBServer. The objects have a pointer to the underlying
MXS_MONITORED_SERVER. The purpose is to have the monitor mainly use
MariaDBServer instead of the current mix of MXS_MONITORED_SERVER* and
MySqlServerInfo and to simplify the mapping between the two. Also,
many methods can be moved to the MariaDBServer class later on.
Some functions have been converted from MXS_MONITORED_SERVER* to
MariaDBServer.
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.
Exposing the canonicalization code in the luafilter allows it to be used
on the Lua side of things. This should enable some pretty cool stuff to be
done with it.
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.
In theory, the value of m_master could change between reading it to
local variable and stopping monitor. To be on the safe side, stop the
monitor first.
The DCB pointer in the MySQLProtocol struct doesn't appear to be updated
in all cases which causes it to be an unreliable source. As the session
itself is always available and it always has the service pointer properly
set, it should be used instead.
Also removed the dead protocol compression code and replaced the
parameters with the service capability bits.
By making it conditional, we prevent the problems that arise when the
replication protocol is used in combination with the session state change
tracking. In addition to this, it prevents unnecessary work for routers
and filters that don't need it.
1. Move some remaining class data private.
2. Linebreak long lines.
3. Move current master autoselection inside class method.
4. Remove single-use constant #defines.
5. Monitor status is only written inside loop.
The database-level query now only takes rows with either a global
select privileges or non-null database privileges. The table-level
query only accepts non-null databases and no global privileges,
as users with global select are added by the previous section.
Now it is possible to control the soft and hard ttl of the
cache on a session basis. That is, it is possible to use
different TTLs for different SELECTs.
As the TTL is checked at lookup time, it need not be hardwired
when the storage instance is created. With this changed it is
possible to introduce @maxscale.cache.(soft|hard)_ttl user
variables using which a client can control what TTL should be
applied for a particular kind of data, which is requested by
MXS-1475.
In case the entry in the cache can not be used because the hard
TTL has kicked in, we fetch the data and update the cache
irrespected of the value if @maxscale.cache.populate. That way
an entry that once was put in the cache, will remain in the cache
(as long as there is space).
