causal_reads does not work with servers that update their state via
mechanisms other than the standard replication. In practice this is just
another Galera limitation.
The documentation stated that MVCC reads weren't safe with
transaction_replay when in reality they are not safe in general and
transaction_replay has no effect on it.
This commit adds a new parameter that, when enabled, prunes the session
command history to a known length. This makes it possible to keep a
client-side pooled connection open indefinitely at the cost of making
reconnections theoretically unsafe. In practice the maximum history length
can be set to a value that encompasses a single session using the pooled
connection with no risk to session state integrity. The default history
length of 50 commands is quite likely to be adequate for the majority of
use-cases.
Both the replication lag and the message printing state are saved in SERVER,
although the values are mostly used by readwritesplit. A log message is printed
both when a server goes over the limit and when it comes back below.
Because of concurrency issues, a message may be printed multiple times before
different threads detect the new message state.
Documentation updated to explain the change.
The TOC was added only to the long documentation files to make them easier
to navigate. Also modified the headings for Avro and Binlog routers and did
some minor cleanup.
Enabling the feature by default prevents the master connection from dying
during times when there are very little or no writes. Having a modest ping
interval of 300 seconds serves to minimize the amount of extra work that
both MaxScale and the server have to do while still keeping the
connections in good shape.
The causal_reads_timeout default value is too long when considering the
behavioral changes that MXS-2141 introduced. With a 10 second default
value, a result is returned to the client in a reasonable amount of time.
With causal_reads enabled, the query would return with an error if the
slave was not able to catch up to the master fast enough. By automatically
retrying the query on the master, we're guaranteed that a valid result is
always returned to the client.
Added the `transaction_replay_max_size` parameter that controls the
maximum size of a transaction that can be replayed. If the limit is
exceeded, the stored statements are released thus preventing the
transaction from being replayed.
This limitation prevents accidental misuse of the transaction replaying
system when autocommit is disabled. It also allows the user to control the
amount of memory that MaxScale will use.
The transaction retrying behavior is now configurable and documented. The
`transaction_replay` parameter implicitly enables the required
functionality in the router that it needs.
Enabling the session command history but limiting it to a history of 50
commands allows reconnections for sessions that don't change the state too
often.
As pooled connections will exceed this limit quite fast, they are not able
to reconnect to servers once connections are lost. To solve this problem,
the session command history needs a compaction process that removes
redundant history.
As they are deprecated, they should not be mentioned in the examples. Also
moved the router_options documentation under a Legacy Configuration
heading.
With the `allow_master_change` parameter enabled, sessions can start using
a different master node if one is available. This will not prevent
sessions from closing if a write query is received while no master
replacement is available.
The default value for strict_multi_stmt prevents compound statements and
atomic multi-statement commands from being executed without completely
disabling load balancing. As the new default value will have no practical
effect on all correct uses of readwritesplit, this is a relatively safe
thing to change.
The new parameter allows the session to be "locked" to the master server
after a stored procedure is called. This will keep the session state
consistent if the stored procedure call modifies the state of the session.
The readwritesplit now sends COM_PING queries to backend servers that have
been idle for too long. The option is configured with the
`connection_keepalive` parameter.
The combination of the default values of `disable_sescmd_history=false`
and `max_slave_connections=100%` does not make sense as it is not possible
to find a replacement slave in case an active one fails.
The max_slave_replication_lag parameter for readwritesplit only works for
monitors that detect replication lag. As the MySQL monitor is the only one
that implements this functionality, the parameter only has meaning when
used with master-slave clusters.
The parameters that readwritesplit uses now use the new system. This
removes the need for the qualified parameter processing found in config.c.
All values for router_options are now also accepted as parameters. The
router_options is deprecated and support for it will be removed in a
future version.
If a slave fails while a non-critical read is being executed, the read is
retried on a different server. This is controlled by the new
`retry_failed_reads` option.
Only selects done that are done outside of a transaction and with
autocommit enabled are retried.
The change in readwritesplit routing priorities, where hints have the
highest priority, gives users more options to control how readwritesplit
acts.
For example, this allows read-only stored procedures to be routed to
slaves by adding a hint to the query:
CALL myproc(); -- maxscale route to slave
The readwritesplit documentation also warns the user not to use routing
hints unless they can be absolutely sure that no damage will be done.
