Commit Graph

5313 Commits

Author SHA1 Message Date
699b3909f7 Set current_dcb to NULL if the current DCB is deleted 2017-09-06 13:45:39 +03:00
c542010e67 Fix internal test failures
The adminusers test did not properly initialize all subsystems in
MaxScale. The polling and DCB tests weren't updated with the changes to
the DCB closing.
2017-09-06 13:20:28 +03:00
795722506b MXS-1387: check if MariaDB 10 Slave has GTID request set.
MXS-1387: check if the connecting MariaDB 10 Slave has the GTID request
set and return error code which stops replication.
2017-09-06 11:13:15 +02:00
3770b4da95 Name read/fake queue consistently
Now dcb_readqueue and dvb_fakequeue are named readq and fakeq
respectively, to be consistent with the naming of the write
and delay queue.
2017-09-06 11:30:24 +03:00
84300c6d97 Do not manipulate read queue directly
Protocol modules should not manupulate the read queue directly,
but always access it via the functions created for that purpose.
2017-09-06 11:21:47 +03:00
24044a7376 Add dcb functions
dcb_readq_append()
dcb_readq_prepend()
dcb_readq_set()
dcb_readq_has()
dcb_readq_release()
dcb_readq_get()
dcb_readq_length()

No code but for DCB code itself should directly manipulate the
internals of a DCB. Thesse functions will be taken into use in
protocol modules.
2017-09-06 11:12:32 +03:00
186ee31abf Fix setting of this_thread.current_dcb
Also in the case of fake events, the current dcb must be set.
2017-09-05 16:04:59 +03:00
d3f4723c81 Do not report events for closed dcb
Before each event handler is called, it is checked whether the
dcb has been closed. If it has been, then the event handler is
not called.

The check has to be made before each event handler, because any
event handler can close the dcb.
2017-09-05 13:20:22 +03:00
d931787e2e Use _exit when daemonizing the process
The original process should use _exit if the forking of the child process
is successful. This makes sure that the exit handlers are called only when
the daemon process exits.
2017-09-04 17:56:34 +03:00
ff94a39038 Binlog server: fix closing slave connections called by MySQL admin connection
Binlog server: fix closing slave connections called by MySQL admin
connection
2017-09-04 15:23:24 +02:00
4c5f0d184c Binlog server: force fake Client DCB close in errorReply
Binlog server: force fake Client DCB close in errorReply. This fixes an
issue when START SLAVE, with new config, is called after a previous one
with errors.
2017-09-04 15:23:24 +02:00
2da7a93473 Destroy workers after services
The workers need to be destroyed only after services have been
to ensure that they are around in case the destruction of services
involves the closing of dcbs.
2017-09-04 16:01:42 +03:00
2f0292fd68 Relax dcb_close assert
If the current worker id is -1, we do not insist that the dcb
is closed by the owning thread. That will happen only for dcbs
that are created before the workers have been started and hence
it is also ok to close them after the workers have exited.
2017-09-04 16:00:27 +03:00
b43ab674e3 Add dcb_close_in_owning_thread(DCB*);
Allows a DCB to be closed by a thread other than the owning
thread.
2017-09-04 13:41:16 +03:00
ccbff0f6d4 Allow debugging of dcb event handling
Using DCB_LOG_EVENT_HANDLING is defined, a notice will be
logged for each event when it is handled.
2017-09-04 11:29:18 +03:00
1a468049b8 Overwrite dcb thread id at final close 2017-09-04 11:09:48 +03:00
0a1608d90d Log proper message in errorReply when master state is BLRM_SLAVE_STOPPED
Log proper message in errorReply when master state is BLRM_SLAVE_STOPPED
2017-09-01 16:48:20 +02:00
85f395394b Connection retry limit minor changes
Connection retry limit minor changes
2017-09-01 13:52:40 +02:00
9230420872 The blr_master_close() is now called in the main worker thread.
The blr_master_close() is now called in the main worker thread by:
blr_stop_slave() and blr_stop_start_master()
2017-09-01 11:49:27 +02:00
18cb61f7de Delay closing in dcb_close
If a dcb being closed is the dcb for which events are currently being
processed, the dcb is not closed immediately but only after all events
have been delivered.
2017-09-01 11:03:23 +03:00
cfe06a563e Remove saving/updating GTID repo when maxscale starts
Remove saving/updating GTID repo when maxscale starts.
The repo update slows down the maxscale startup.
The saving was added in order populate GTID repo with current file but
it costs extra time to event validation routine when current binlog has
many events and it has a big size.
2017-08-31 17:27:52 +02:00
d7b04fc451 Add static storage to anonymous unit/thread structures
Without static, CentOS6 complains about non-local variable
using anonymous type.
2017-08-31 12:45:16 +03:00
e1efb91ea5 MXS-1156: MASTER_CONNECT_RETRY is a new option for CHANGE MASTER TO
MXS-1156: MASTER_CONNECT_RETRY is a new option for CHANGE MASTER TO.

The ‘connect_retry’, ‘master_retry_count’ options are now in use for
both service and
router_options.
2017-08-31 11:01:03 +02:00
42833386f5 MXS-1156: added configuration entry point for connection retry parameters.
MXS-1156: added configuration entry points for connection retry
parameters.

Variable names have been also changed.
2017-08-31 11:01:03 +02:00
c91187d4de MXS-1156: added limit to master connect retry
MXS-1156: added limit to master connect retry
2017-08-31 11:01:03 +02:00
3c7a0014a9 MXS-1156: Store master_heartbeat_period in master.ini
MXS-1156: Store master_heartbeat_period in master.ini
2017-08-31 11:01:03 +02:00
0f5f9f8228 MXS-1156: Disable heartbeat if value is set to 0
MXS-1156: Disable heartbeat if value is set to 0.

A warning log is written in maxscale log file when heartbeat is not set.
2017-08-31 11:01:03 +02:00
5b2cfac3cb MXS-1156: Added new option MASTER_HEARTBEAT_PERIOD to CHANGE MASTER TO
MXS-1156: Added new option MASTER_HEARTBEAT_PERIOD to CHANGE MASTER TO
2017-08-31 11:01:03 +02:00
68b7d88fa1 Check return value of freopen
Otherwise compilation fails in release mode.
2017-08-30 14:36:51 +03:00
63257f1b2a MXS-1378 Log session id
If the session id is known, it will be logged together with all
messages. If present, the session id appears, enclosed in paranthesis,
right after the message category. E.g.

    2017-08-30 12:20:49   warning: (4711) [masking] The rule ...
2017-08-30 13:58:42 +03:00
e7926d770e MXS-1378 Provide access to current session
Provide access to current session and session id. This will
be used by the logging mechanism for logging the session id
together with messages.
2017-08-30 13:58:42 +03:00
67efd1daea MXS-1378 Provide access to the current DCB
This will be used by a subsequent `session_get_current()` and
`session_get_current_id()` for obtaining the current SESSION and
session id, respectively. The latter of those will be used by the
logging mechanism for logging the session id in conjunction with
messages.
2017-08-30 13:58:42 +03:00
e6f78bfcb3 Pass modutil function state as a struct
This allows the type to change without changing it at the call sites.
2017-08-29 22:46:35 +03:00
2a346fd061 MXS-1376 Really close a DCB only at the end of event loop
When dcb_close() is called, the DCB is only marked for closing
and the actual closing takes place only after all event handlers
have been called. That way, the state of the DCB will not change
during event processing but only after.

From a handler perspective this should now be just like it was
when the zombie queue was present.

TODO: There are far too many state variables or variables akin to
      state variables - dcb_role, state, persistentstart, n_close -
      in DCB. A cleanup is warranted.
2017-08-29 15:44:50 +03:00
13f7015e7b Fix EOF packet calculation for large rows
The EOF packet calculation function in modutil.cc didn't handle the case
where the payload exceeded maximum packet size and could mistake binary
data for a ERR packet.

The state of a multi-packet payload is now exposed by the
modutil_count_signal_packets function. This allows proper handling of
large multi-packet payloads.

Added minor improvements to mxs1110_16mb to handle testing of this change.
2017-08-25 17:00:42 +03:00
9c25e6d995 MXS-1376 All zombie related code removed
As dcbs are now closed when dcb_close() is called and there is
no zombie queue, the zombie state can also be removed.
2017-08-25 14:48:16 +03:00
c11d8fa328 MXS-1376 Close a connection in dcb_close
As each client/server connection will be handled by a specific
thread, all closing activity can take place directly when the
connection is closed and not later when the zombie queue is
processed.

In a subsequent commit the zombie queue will be removed.
2017-08-25 14:48:16 +03:00
c2a9c4bfc1 Create admin user files if they don't exist
The default users are now inserted into the admin users files if no
existing files are found. This removes the hard-coded checks for admin
user names and simplifies the admin user logic.
2017-08-25 13:54:10 +03:00
0e89e445f2 Bind admin interface to IPv4 if IPv6 fails
The default interface for the admin interface is the IPv6 address '::'
which corresponds to the IPv4 address '0.0.0.0'. If the system doesn't
support IPv6, then an attempt to bind on IPv4 should be made.
2017-08-25 13:54:10 +03:00
894ef3d3a9 MXS-1346: Compile dbfwfilter as C++
Compile the dbfwfilter as C++ to allow STL container usage.
2017-08-25 13:54:10 +03:00
8927298abe MXS-1075: Avoid loops when missing binlog file is detected.
MXS-1075: Avoid loops when missing binlog file is detected.
2017-08-24 16:03:30 +02:00
ad4e8dad94 MXS-1364 Drop the usage field
But for the most trivial statements did not really provide
useful information.

The arguments of the "function" '=' are now reported.
2017-08-24 09:58:18 +03:00
72831c86ac DCBEVENTQ removed
No longer used for anything
2017-08-24 09:57:42 +03:00
11dffd0d93 Remove unused variables 2017-08-24 09:57:42 +03:00
c05bb3593a MXS-1346: Take the new query classifier features into use
The function usage by column is now used by the dbfwfilter. This finishes
the implementation of blocking function usage per column.
2017-08-24 09:46:03 +03:00
da74ebb959 Add error string helper function to dbfwfilter
The helper function is a wrapper for snprintf and strdup which removes
some of the redundant code in dbfwfilter.
2017-08-24 09:46:02 +03:00
ae2b9fd30d Fix hang with queries larger than 2^24 bytes
Readwritesplit didn't track multi-packet queries which resulted in them
being confused for pipelined queries.
2017-08-24 09:46:02 +03:00
2ef9fbc47a MXS-1346: Add uses_function type rule
The uses_function type rule matches when any of the columns given as
values uses a function. With this, columns can be denied from being used
with a function.
2017-08-24 09:46:02 +03:00
0aa1b8a4d0 MXS-1344: MySQL monitor detects MaxScale Binlog Server
If MaxScale Binlog Server is detected by MySQL monitor, then Server
status is ‘Relay Master, Running’ without ‘Slave’: this way Binlog
Server cannot be used by any statement routing.
Additionally SELECT for replication lag is skipped as well.
2017-08-23 14:11:17 +02:00
1c4ae9f89a MXS-1344: Allow SET SQL_MODE='' statement from mysql monitor
When SET SQL_MODE='' statement is received from mysql monitor an OK
packet is sent.
2017-08-23 14:11:16 +02:00