Commit Graph

5814 Commits

Author SHA1 Message Date
3e85500491 Merge branch '2.3' into 2.4 2019-07-02 08:38:15 +03:00
4954c7f6b7 Fix sending of unknown PS error
The error was only generated for COM_STMT_EXECUTE commands when all PS
commands should trigger it. In addition, large packets would get sent two
errors upon the arrival of the trailing end.
2019-07-01 10:25:41 +03:00
40f35618ad MXS-2562: Fix out-of-order error during COM_CHANGE_USER
If an error is generated while a COM_CHANGE_USER is being done, it would
always use the sequence number 1. To properly handle this case and send
the correct sequence number, the COM_CHANGE_USER progress needs to be
tracked at the session level.

The information needs to be shared between the backend and client
protocols as the final OK to the COM_CHANGE_USER, with the sequence number
3, is the one that the backend server returns. Only after this response
has been received and routed to the client can the COM_CHANGE_USER
processing stop.
2019-06-28 20:25:31 +03:00
3b6387c952 MXS-2562: Stop immediately on mid-resultset failure
If a server fails mid-resultset, there's not a lot we can do to recover
the situation. A few cases could be handled (e.g. generate an ERR if the
resultset has proceeded to the row processing stage) but these fall
outside the scope of the original issue.
2019-06-28 20:25:31 +03:00
74d15cee09 MXS-2562: Always send errors with sequence number 1
As is explained in MDEV-19893:

   Client reads from socket, gets the packet from 1. with seqno=0, which
   it does not expect, since seqno is supposed to be incremented. Client
   complains, throws tantrums and exceptions.

To cater for clients that do not expect out-of-bound messages
(i.e. server-initiated packets with seqno 0), all messages generated by
MaxScale should use at least sequence number 1.
2019-06-28 20:25:31 +03:00
5dca53f877 Merge branch '2.4.0' into 2.4 2019-06-28 17:58:01 +03:00
5ab5e914e7 MXS-2582: Deep-copy PS buffers in RWBackend::write
Deep-copying prevents subsequent modifications done by the caller from
affecting the data that can be potentially stored in the write queue of
the backend's DCB.
2019-06-28 09:46:15 +03:00
3b39d42d3b Merge branch '2.3' into 2.4 2019-06-27 18:56:25 +03:00
166d26ff13 Avoid using SQLITE_OPEN_URI
Centos6 uses a very old version of SQLite without support for URI filenames.
PAM authenticator must use a file-based database.

Commit cherry-picked to 2.4.0 from 2.3.
2019-06-27 15:18:49 +03:00
d8790fa3e7 Merge branch '2.2' into 2.3 2019-06-27 14:20:42 +03:00
b4289224de Avoid using SQLITE_OPEN_URI
Centos6 uses a very old version of SQLite without support for URI filenames.
PAM authenticator must use a file-based database.
2019-06-27 13:59:05 +03:00
8a749dc627 MXS-2521: Fix build failure
The get_target_type signature for the definition was different from the
declaration.
2019-06-27 12:48:03 +03:00
1d87b88c50 Avoid using sqlite3_errstr()
The function is not in old SQLite3-versions such as the one used by
Centos6.
2019-06-27 10:55:24 +03:00
fcc19f3c66 Merge branch '2.3' into 2.4 2019-06-27 08:53:33 +03:00
5eba688c1b MXS-2521: Detect COM_STMT_EXECUTE without metadata
If a COM_STMT_EXECUTE has no metadata in it and it has more than one
parameter, it must be routed to the same backend where the previous
COM_STMT_EXECUTE with the same ID was routed to. This prevents MDEV-19811
that is triggered by MaxScale routing the queries to different backends.
2019-06-27 08:29:30 +03:00
5c762bb841 misc fix 2019-06-27 08:17:42 +03:00
51ce3c53fd shift 32 2019-06-27 08:17:42 +03:00
6b31b80e76 fix wrong assert 2019-06-27 08:17:42 +03:00
8d50450b5a MXS-2521:Route subseqenct COM_STMT_EXECUTE to the same server which first COM_STMT_EXECUTE was executed on 2019-06-27 08:17:42 +03:00
25a076d072 Fix buffer sizes in PS extraction code
The code used 4 byte buffers for 2 byte values.
2019-06-26 15:49:50 +03:00
d45915c279 Treat empty avro filenames as an error
If an empty filename is requested, treat it as an error.
2019-06-26 10:17:21 +03:00
01822490ca Fix avrorouter and cluster usage
The servers defined by `cluster` wouldn't get used.
2019-06-26 10:17:21 +03:00
59e7b2b09d Deprecate mqfilter 2019-06-26 10:08:15 +03:00
d41d4021a8 Merge branch '2.3' into 2.4.0 2019-06-26 09:20:35 +03:00
67d82146f1 Fix packet length assertion
The `used_len + len` can be the same as `total_len`.
2019-06-26 09:19:24 +03:00
88dae197b0 Make packet boundary assertion less costly
mxs::Buffer::iterator is not a random-access iterator and hence will have
linear cost. This is too costly to do on every packet with even moderately
sized resultsets.
2019-06-26 09:13:03 +03:00
f961bf1d96 Merge branch '2.3' into 2.4.0 2019-06-25 11:46:44 +03:00
b8bb8702bd Fix broken Change Dates
Were not correct, so were not updated with all others.
2019-06-25 09:52:42 +03:00
2ab9aa9a94 Update 2.4.0 Change Date 2019-06-25 09:19:55 +03:00
873ea92915 Assert that packet is inside the buffer
The buffer containing the response should contain only complete packets.
2019-06-25 09:05:49 +03:00
9f771ffc2b Revert "MXS-2513 Deprecate firewall and masking filters"
This reverts commit de3c5721e95d5d5be909a56faeb121ad09a2f6fe.
2019-06-24 10:21:01 +03:00
278c26cd54 Fix build failure
mxs::rtrim was renamed to mxb::rtrim.
2019-06-20 23:17:41 +03:00
0b67ce1e82 Merge branch '2.3' into develop 2019-06-20 14:36:48 +03:00
be429c2c57 MXS-2490: Always restore original statement ID
By always restoring the ID, we are guaranteed to only store the query in
the form that it was originally sent in. This should be changed so that
the ID that the client sends can be used as-is in the backends.
2019-06-20 14:27:04 +03:00
805be70a78 Add more information to rwsplit info messages
The statement ID for all binary protocol statements and the error given to
handleError are now logged.
2019-06-20 14:27:03 +03:00
20f595d978 Merge branch '2.3' into develop 2019-06-20 13:00:13 +03:00
301b1b63ab MXS-2569: Always terminate schemas with newlines
If the stored file didn't have a newline in the schema, the schema and the
first row would be on the same line.
2019-06-20 12:32:30 +03:00
197d577aa1 Set correct replication lag state
The lag state was inverted.
2019-06-19 17:35:19 +03:00
5528a8e8fb Merge branch '2.3' into develop 2019-06-19 17:27:57 +03:00
39df829384 Don't kick out super-users when performing rejoin
When rejoining a server, preventing writes is not as essential as in switchover.
2019-06-19 17:14:39 +03:00
2bd614ce8e MXS-2566: Use connection counts more often
The connection counts are now always used to pick the best servers where
the initial connections are created. This covers both master and slaves
selection. Reconnections done while routing queries still pick the "best"
server according to the slave selection criteria. This allows better
servers to be taken into use when `lazy_connect` is enabled.
2019-06-19 16:12:28 +03:00
ef3136c5c9 MXS-2555 Fix split packet bug in smart session.
The test failed to notice that the split request session was left in a
bad state after the split query had (succesfully) finished.
2019-06-19 15:31:34 +03:00
6ba1c277be MXS-2555 Fix a few TODOs and other cleanup 2019-06-19 15:31:34 +03:00
a950715176 MXS-2555 Implement cache eviction (or query re-measurment) strategy 2019-06-19 15:31:34 +03:00
f5615cc174 MXS-2555 Add support for unix sockets to LocalClient 2019-06-19 15:31:34 +03:00
f7d44835a7 MXS-2555 Use mxs_mysql_execute_kill_all_others in smartrouter 2019-06-19 15:31:34 +03:00
2343326ecc MXS-2555 Add execute_kill version, that kills all but a given mysql_thread_id
NOTE: This does not work whit unix scokets. The reason is that LocalClient
assumes network sockets. TODO: to add unix socket support.
2019-06-19 15:31:34 +03:00
56e5c417c1 MXS-2555 uncrustify mysql_common.cc, and a minor bug fix in smartrouter.cc 2019-06-19 15:31:34 +03:00
6a7b6d4b89 MXS-2555 Beta Smart Router.
1. Remove persistence of performance data
2. Move global CanonicalPerformance into SmartRouter object
3. Implement another kill_all_others_v2. Left kill_all_others_v1
    in case it should be fixed and used instead.
2019-06-19 15:31:34 +03:00
1241300494 MXS-2555 Simple Smart Router
Does the measurments, usage of the same  and persistence to file.
Kill is not implemented, so waits for all clusters to fully responds

The performance data uses a mutex and the persistence data file
is written while holding the mutex. This obviously needs to be
improved, but this commit shows the working concept.
2019-06-19 15:31:34 +03:00