Commit Graph

453 Commits

Author SHA1 Message Date
Amber Brown b69216f768
Make the metrics less racy (#4061) 2018-10-19 21:45:45 +11:00
Richard van der Hoff b8a5b0097c
Various cleanups in the federation client code (#4031)
- Improve logging: log things in the right order, include destination and txids
  in all log lines, don't log successful responses twice

- Fix the docstring on TransportLayerClient.send_transaction

- Don't use treq.request, which is overcomplicated for our purposes: just use a
  twisted.web.client.Agent.

- simplify the logic for setting up the bodyProducer

- fix bytes/str confusions
2018-10-16 10:44:49 +01:00
Erik Johnston 334e075dd8 Fix error when logging incomplete requests
If a connection is lost before a request is read from Request, Twisted
sets `method` (and `uri`) attributes to dummy values. These dummy values
have incorrect types (i.e. they're not bytes), and so things like
`__repr__` would raise an exception.

To fix this we had a helper method to return the method with a
consistent type.
2018-10-02 12:06:22 +01:00
Richard van der Hoff 94f7befc31
Merge pull request #3925 from matrix-org/erikj/fix_producers_unregistered
Fix spurious exceptions when client closes conncetion
2018-09-25 11:52:06 +01:00
Erik Johnston b28a7ed503 Fix spurious exceptions when client closes conncetion
If a HTTP handler throws an exception while processing a request we
automatically write a JSON error response. If the handler had already
started writing a response twisted throws an exception.

We should check for this case and simple abort the connection if there
was an error after the response had started being written.
2018-09-20 13:44:20 +01:00
Amber Brown 1f3f5fcf52
Fix client IPs being broken on Python 3 (#3908) 2018-09-20 20:14:34 +10:00
Amber Brown aeca5a5ed5
Add a regression test for logging on failed connections (#3912) 2018-09-20 16:28:18 +10:00
Erik Johnston a334e1cace Update to use new timeout function everywhere.
The existing deferred timeout helper function (and the one into twisted)
suffer from a bug when a deferred's canceller throws an exception, #3842.

The new helper function doesn't suffer from this problem.
2018-09-19 10:39:40 +01:00
Amber Brown 47c02e6332
Merge pull request #3909 from turt2live/travis/fix-logging-1
Fix matrixfederationclient.py logging: Destination is a string
2018-09-19 18:14:47 +10:00
Amber Brown 3d6b24fb1b
Merge pull request #3907 from matrix-org/rav/set_sni_to_server_name
Set SNI to the server_name, not whatever was in the SRV record
2018-09-19 17:59:33 +10:00
Travis Ralston 35aec19f0a Destination is a string 2018-09-18 15:29:30 -06:00
Richard van der Hoff 31c15dcb80
Refactor matrixfederationclient to fix logging (#3906)
We want to wait until we have read the response body before we log the request
as complete, otherwise a confusing thing happens where the request appears to
have completed, but we later fail it.

To do this, we factor the salient details of a request out to a separate
object, which can then keep track of the txn_id, so that it can be logged.
2018-09-18 18:17:15 +01:00
Richard van der Hoff b3097396e7 Set SNI to the server_name, not whatever was in the SRV record
Fixes #3843
2018-09-18 17:01:12 +01:00
Richard van der Hoff ac80cb08fe Fix more b'abcd' noise in metrics 2018-09-17 17:16:50 +01:00
Richard van der Hoff f00a9d2636 Fix some b'abcd' noise in logs and metrics
Python 3 compatibility: make sure that we decode some byte sequences before we
use them to create log lines and metrics labels.
2018-09-17 16:15:42 +01:00
Erik Johnston fcfe7a850d Add an awful secondary timeout to fix wedged requests
This is an attempt to mitigate #3842 by adding yet-another-timeout
2018-09-14 19:23:07 +01:00
Erik Johnston 3e6e94fe9f
Merge pull request #3872 from matrix-org/hawkowl/timeouts-2
timeouts 2: electric boogaloo
2018-09-14 16:58:44 +01:00
Amber Brown bc9af88a2d fix 2018-09-15 00:26:00 +10:00
Erik Johnston 9e2f9a7b57 Measure outbound requests 2018-09-14 15:11:26 +01:00
Amber Brown c971aa7b9d fix 2018-09-14 03:57:02 +10:00
Amber Brown 8f08d848f5 fix 2018-09-14 03:53:56 +10:00
Amber Brown 7c33ab76da redact better 2018-09-14 03:45:34 +10:00
Amber Brown 63755fa4c2 we do that higher up 2018-09-14 03:21:47 +10:00
Amber Brown 73884ebac5 Merge remote-tracking branch 'origin/develop' into hawkowl/timeouts-2 2018-09-14 03:11:25 +10:00
Amber Brown 7c27c4d51c
merge (#3576) 2018-09-14 03:11:11 +10:00
Amber Brown 1c3f4d9ca5 buffer? 2018-09-14 03:09:13 +10:00
Amber Brown bfa0b759e0
Attempt to figure out what's going on with timeouts (#3857) 2018-09-14 00:15:51 +10:00
Erik Johnston 8c5b84441b Log outbound requests when we retry 2018-09-12 16:22:14 +01:00
Erik Johnston 4084a774a8 Timeout reading body for outbound HTTP requests 2018-09-12 10:10:20 +01:00
Amber Brown 2d2828dcbc
Port http/ to Python 3 (#3771) 2018-09-06 00:10:47 +10:00
Erik Johnston 1d67b13674 Fix bug when federation_domain_whitelist is an emtpy list
Outbound federation were incorrectly allowed when the config option was
set to an empty list
2018-08-24 12:13:12 +01:00
Richard van der Hoff a52f276990 Bugfixes
--------
 
 - Fix bug in v0.33.3rc1 which caused infinite loops and OOMs
 ([\#3723](https://github.com/matrix-org/synapse/issues/3723))
 -----BEGIN PGP SIGNATURE-----
 
 iQEzBAABCAAdFiEEIQBQJ4l+yK4dlKkFIwi0edOSShEFAlt6/yMACgkQIwi0edOS
 ShFrzAgAhLA74fClydGLfhPNCXIw12Amv2ASA6+r0ukqjYQ/9MZ0CRsL+PMZJTk6
 Z75PK3TkX38zSpPaUx9HdMQO/yE+rpI99ZxlBM+8jzsf1jyF8ljqxSfixXUlaXX9
 698PQQ3f+IDoSs59NKrHvoWhUThaBpf/1GwRxG5mUcm2oKJItNMHk/uLurNCaIIZ
 YSwWcxvZdluOMX0WrlF/S524DqtgULcRf3E9x3mumobkVGxI3UH73hXgGUcPi7XV
 ttLuy7fVbV2Cun/wvjBqD7kc8LNbKwSaT9y+/xS0D5QYjph4rpQiTG7S5rtPLPMU
 Z2lL1C5ggBCkD43g3orDvehbNWCL5g==
 =0vEe
 -----END PGP SIGNATURE-----

Merge tag 'v0.33.3rc2' into develop

Bugfixes
--------

- Fix bug in v0.33.3rc1 which caused infinite loops and OOMs
([\#3723](https://github.com/matrix-org/synapse/issues/3723))
2018-08-21 20:30:09 +01:00
Erik Johnston cd6937fb26 Fix typo 2018-08-21 16:28:10 +01:00
Richard van der Hoff be6527325a Fix exceptions when a connection is closed before we read the headers
This fixes bugs introduced in #3700, by making sure that we behave sanely
when an incoming connection is closed before the headers are read.
2018-08-20 18:21:10 +01:00
Erik Johnston b01a755498 Make the in flight requests metrics thread safe 2018-08-20 17:27:52 +01:00
Amber Brown 324525f40c
Port over enough to get some sytests running on Python 3 (#3668) 2018-08-20 23:54:49 +10:00
Richard van der Hoff 63260397c6
Merge pull request #3701 from matrix-org/rav/use_producer_for_responses
Use a producer to stream back responses
2018-08-17 14:58:45 +01:00
Richard van der Hoff afcd655ab6 Use a producer to stream back responses
The problem with dumping all of the json response into the Request object at
once is that doing so starts the timeout for the next request to be received:
so if it takes longer than 60s to stream back the response to the client, the
client never gets it.

The correct solution is to use a Producer; then the timeout is only started
once all of the content is sent over the TCP connection.
2018-08-15 15:04:16 +01:00
Richard van der Hoff eaaa2248ff Refactor request logging code
This commit moves a bunch of the logic for deciding when to log the receipt and
completion of HTTP requests into SynapseRequest, rather than in the request
handling wrappers.

Advantages of this are:
 * we get logs for *all* requests (including OPTIONS and HEADs), rather than
   just those that end up hitting handlers we've remembered to decorate
   correctly.

 * when a request handler wires up a Producer (as the media stuff does
   currently, and as other things will do soon), we log at the point that all
   of the traffic has been sent to the client.
2018-08-15 13:47:52 +01:00
Amber Brown b37c472419
Rename async to async_helpers because `async` is a keyword on Python 3.7 (#3678) 2018-08-10 23:50:21 +10:00
Richard van der Hoff 3c0213a217
Merge pull request #3439 from vojeroen/send_sni_for_federation_requests
send SNI for federation requests
2018-08-10 12:23:54 +01:00
Richard van der Hoff f900d50824 include known room versions in outgoing make_joins 2018-08-06 13:45:37 +01:00
Richard van der Hoff 01e93f48ed Kill off MatrixCodeMessageException
This code brings the SimpleHttpClient into line with the
MatrixFederationHttpClient by having it raise HttpResponseExceptions when a
request fails (rather than trying to parse for matrix errors and maybe raising
MatrixCodeMessageException).

Then, whenever we were checking for MatrixCodeMessageException and turning them
into SynapseErrors, we now need to check for HttpResponseExceptions and call
to_synapse_error.
2018-08-01 16:02:46 +01:00
Richard van der Hoff fa7dc889f1 Be more careful which errors we send back over the C-S API
We really shouldn't be sending all CodeMessageExceptions back over the C-S API;
it will include things like 401s which we shouldn't proxy.

That means that we need to explicitly turn a few HttpResponseExceptions into
SynapseErrors in the federation layer.

The effect of the latter is that the matrix errcode will get passed through
correctly to calling clients, which might help with some of the random
M_UNKNOWN errors when trying to join rooms.
2018-08-01 16:02:38 +01:00
Amber Brown da7785147d
Python 3: Convert some unicode/bytes uses (#3569) 2018-08-02 00:54:06 +10:00
Richard van der Hoff 7041cd872b
Merge branch 'develop' into send_sni_for_federation_requests 2018-07-27 09:17:11 +01:00
Amber Brown a97c845271
Move v1-only APIs into their own module & isolate deprecated ones (#3460) 2018-07-19 20:03:33 +10:00
Erik Johnston 5bd0a47fcd pep8 2018-07-18 14:19:00 +01:00
Erik Johnston e45a46b6e4 Add response code to response timer metrics 2018-07-18 13:59:36 +01:00
Amber Brown bc006b3c9d
Refactor REST API tests to use explicit reactors (#3351) 2018-07-17 20:43:18 +10:00