The tests are now built by default. This should make it easier for users
to verify that they have a working MaxScale.
Also made the building of test_parse_kill conditional like the rest of the
tests.
When four servers (A, B, C and E where E and A replicate from each other
and A is the master for B and C) form a cluster and only three of them (A,
B and C) are configured into MaxScale, a failover operation from A to B
(making B the current master) and a restart of A causes B to lose its
master status.
The following diagram illustrates the state of the cluster at the end of
the process described above.
+----------------------+
| +---+ |
+------------+ B <-+ |
+-v-+ | +---+ | |
| E | | | |
+-^-+ | +---+ +-+-+ |
+------+ A | | C | |
| +---+ +---+ |
| |
+----------------------+
The external server E was not correctly ignored in the replication
topology generation causing both A and B to be seen as the lowest slave
nodes in the tree. From a theoretical point of view this is the correct
interpretation as there are two distinct trees and neither of them
contains any true masters.
In practice, MaxScale should treat any servers that replicate from an
external master as root level master nodes. Doing this guarantees that they
are labeled as masters if they have slaves replicating from them.
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.
qc_mysqlembedded must include the headers from the embedded
library and not from Connector-C.
So, it must be ensured the embedded library include directories
are first.
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 MariaDB Connector-C headers that are built by MaxScale must be
included before any system headers.
Fixed code that explicitly included the <mysql.h> header to use the
<maxscale/protocol/mysql.h> wrapper instead.
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 test checks that failover works even when the master of the monitored
cluster is a slave to an external masters. The test also verifies that the
servers do not get unexpected status labels.
When detect_standalone_master is enabled, the root_master variable was not
updated after the master was changed by the standalone server detection
mechanism. This caused debug assertions to fire in addition to possibly
causing some of the ignore_external_masters logic to break.
"servers_no_promotion" is a comma-separated list of servers
which cannot be chosen when selecting a new master during failover
(auto or manual), or when automatically selecting a new master
for switchover (currently disabled).
The servers in the list are redirected normally and can be promoted
by switchover when manually selecting a new master.
If the API versions do not match, MaxScale will treat this as an
error. The API versioning would allow backwards compatible changes but the
functionality to handle that is not implemented in MaxScale.
Updated API versions based on changes done to module APIs in 2.2.
With the flag --debug=enable-statement-logging it is now possible
to instruct MaxScale to log all SQL statements it sends to the
servers.
The format of the logged string looks like:
notice : SQL(127.0.0.1): 0, "SELECT ..."
First the fixed string "SQL", followed by the server address in
parenthesis followed by the actual return value of mysql_query(),
followed by the statement itself.
The "SQL" string makes the lines easy to grep for and having the
return value before the statement makes it easier to spot since
the length of the return value string does not wary much, but the
length of the statements do wary a lot.
The Master status now prevents Slave status from being assigned to a
server. In practice this simply means that the master will not have both
the Master and Slave status bits.
The test uses a separate writer thread to insert data into the
master. This thread must be halted before the blocking of the master
happens as the slaves must catch up. Once slaves have caught up and the
master is blocked, the writer thread can continue doing inserts.
At the end of the test slaves must also be synchronized before the
inserted data is validated. This prevents test failures due to slave lag.
Tests that local_address is taken into account. However, at the time
of writing the maxscale VM does not have two usable IP addresses, so
we only test that explicitly specifying an IP-address does not break
things.
Locally it has been confirmed that this indeed works the way it is
supposed to.
In some cases you might want to use a specific address/interface
when connecting to a server instead of the default one. With the
global parameter 'local_address' it can now be specified which
address to use.
The test repeatedly blocks and unblocks a master which goes unnoticed by
the monitor due to the 10 second read timeouts in the monitor
configuration in the longblob template. The replication template uses the
default timeouts which makes the test actually exercise the functionality
that it is intended to test.
By always starting the session shutdown process by stopping the client
DCB, the manipulation of the session state can be removed from the backend
protocol modules and replaced with a fake hangup event.
Delivering this event via the core allows the actual dcb_close call on the
client DCB to be done only when the client DCB is being handled by a
worker.
Directly closing the client DCB in the backend protocol modules is not
correct anymore as the state of the session doesn't change when the client
DCB is closed. By propagating the shutdown of the session with a fake
hangup to the client DCB, the closing of the DCB is done only once.
Added debug assertions that make sure all DCBs are closed only
once. Removed redundant code in the backend protocol error handling code.
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.
The test now also creates TIME type values and checks that they are
converted correctly. Also added NULL value tests for all values and made
required adjustments to the code.
In debug mode, when scanning the server id from a string, check that resulting
number is 32bit. Also, when querying the server id, query the global version.
Now, if a super user modifies the server id the monitor will notice it.
Server id:s in gtid:s are handled similarly.
When a double close is detected in a debug build, a debug assertion is
triggered. This will generate a core dump which should help investigate
the double close.
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.
