Mercurial > hg-stable
view tests/test-http-api-httpv2.t @ 37288:9bfcbe4f4745
wireproto: add streams to frame-based protocol
Previously, the frame-based protocol was just a series of frames,
with each frame associated with a request ID.
In order to scale the protocol, we'll want to enable the use of
compression. While it is possible to enable compression at the
socket/pipe level, this has its disadvantages. The big one is it
undermines the point of frames being standalone, atomic units that
can be read and written: if you add compression above the framing
protocol, you are back to having a stream-based protocol as opposed
to something frame-based.
So in order to preserve frames, compression needs to occur at
the frame payload level.
Compressing each frame's payload individually will limit compression
ratios because the window size of the compressor will be limited
by the max frame size, which is 32-64kb as currently defined. It
will also add CPU overhead, as it is more efficient for compressors
to operate on fewer, larger blocks of data than more, smaller blocks.
So compressing each frame independently is out.
This means we need to compress each frame's payload as if it is part
of a larger stream.
The simplest approach is to have 1 stream per connection. This
could certainly work. However, it has disadvantages (documented below).
We could also have 1 stream per RPC/command invocation. (This is the
model HTTP/2 goes with.) This also has disadvantages.
The main disadvantage to one global stream is that it has the very
real potential to create CPU bottlenecks doing compression. Networks
are only getting faster and the performance of single CPU cores has
been relatively flat. Newer compression formats like zstandard offer
better CPU cycle efficiency than predecessors like zlib. But it still
all too common to saturate your CPU with compression overhead long
before you saturate the network pipe.
The main disadvantage with streams per request is that you can't
reap the benefits of the compression context for multiple requests.
For example, if you send 1000 RPC requests (or HTTP/2 requests for
that matter), the response to each would have its own compression
context. The overall size of the raw responses would be larger because
compression contexts wouldn't be able to reference data from another
request or response.
The approach for streams as implemented in this commit is to support
N streams per connection and for streams to potentially span requests
and responses. As explained by the added internals docs, this
facilitates servers and clients delegating independent streams and
compression to independent threads / CPU cores. This helps alleviate
the CPU bottleneck of compression. This design also allows compression
contexts to be reused across requests/responses. This can result in
improved compression ratios and less overhead for compressors and
decompressors having to build new contexts.
Another feature that was defined was the ability for individual frames
within a stream to declare whether that individual frame's payload
uses the content encoding (read: compression) defined by the stream.
The idea here is that some servers may serve data from a combination
of caches and dynamic resolution. Data coming from caches may be
pre-compressed. We want to facilitate servers being able to essentially
stream bytes from caches to the wire with minimal overhead. Being
able to mix and match with frames are compressed within a stream
enables these types of advanced server functionality.
This commit defines the new streams mechanism. Basic code for
supporting streams in frames has been added. But that code is
seriously lacking and doesn't fully conform to the defined protocol.
For example, we don't close any streams. And support for content
encoding within streams is not yet implemented. The change was
rather invasive and I didn't think it would be reasonable to implement
the entire feature in a single commit.
For the record, I would have loved to reuse an existing multiplexing
protocol to build the new wire protocol on top of. However, I couldn't
find a protocol that offers the performance and scaling characteristics
that I desired. Namely, it should support multiple compression
contexts to facilitate scaling out to multiple CPU cores and
compression contexts should be able to live longer than single RPC
requests. HTTP/2 *almost* fits the bill. But the semantics of HTTP
message exchange state that streams can only live for a single
request-response. We /could/ tunnel on top of HTTP/2 streams and
frames with HEADER and DATA frames. But there's no guarantee that
HTTP/2 libraries and proxies would allow us to use HTTP/2 streams
and frames without the HTTP message exchange semantics defined in
RFC 7540 Section 8. Other RPC protocols like gRPC tunnel are built
on top of HTTP/2 and thus preserve its semantics of stream per
RPC invocation. Even QUIC does this. We could attempt to invent a
higher-level stream that spans HTTP/2 streams. But this would be
violating HTTP/2 because there is no guarantee that HTTP/2 streams
are routed to the same server. The best we can do - which is what
this protocol does - is shoehorn all request and response data into
a single HTTP message and create streams within. At that point, we've
defined a Content-Type in HTTP parlance. It just so happens our
media type can also work as a standalone, stream-based protocol,
without leaning on HTTP or similar protocol.
Differential Revision: https://phab.mercurial-scm.org/D2907
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Mon, 26 Mar 2018 11:00:16 -0700 |
parents | bbea991635d0 |
children | 5fadc63ac99f |
line wrap: on
line source
$ HTTPV2=exp-http-v2-0001 $ MEDIATYPE=application/mercurial-exp-framing-0002 $ send() { > hg --verbose debugwireproto --peer raw http://$LOCALIP:$HGPORT/ > } $ cat > dummycommands.py << EOF > from mercurial import wireprototypes, wireproto > @wireproto.wireprotocommand('customreadonly', permission='pull') > def customreadonly(repo, proto): > return wireprototypes.bytesresponse(b'customreadonly bytes response') > @wireproto.wireprotocommand('customreadwrite', permission='push') > def customreadwrite(repo, proto): > return wireprototypes.bytesresponse(b'customreadwrite bytes response') > EOF $ cat >> $HGRCPATH << EOF > [extensions] > dummycommands = $TESTTMP/dummycommands.py > EOF $ hg init server $ cat > server/.hg/hgrc << EOF > [experimental] > web.apiserver = true > EOF $ hg -R server serve -p $HGPORT -d --pid-file hg.pid $ cat hg.pid > $DAEMON_PIDS HTTP v2 protocol not enabled by default $ send << EOF > httprequest GET api/$HTTPV2 > user-agent: test > EOF using raw connection to peer s> GET /api/exp-http-v2-0001 HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 404 Not Found\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 33\r\n s> \r\n s> API exp-http-v2-0001 not enabled\n Restart server with support for HTTP v2 API $ killdaemons.py $ cat > server/.hg/hgrc << EOF > [experimental] > web.apiserver = true > web.api.http-v2 = true > EOF $ hg -R server serve -p $HGPORT -d --pid-file hg.pid $ cat hg.pid > $DAEMON_PIDS Request to unknown command yields 404 $ send << EOF > httprequest POST api/$HTTPV2/ro/badcommand > user-agent: test > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/badcommand HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 404 Not Found\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 42\r\n s> \r\n s> unknown wire protocol command: badcommand\n GET to read-only command yields a 405 $ send << EOF > httprequest GET api/$HTTPV2/ro/customreadonly > user-agent: test > EOF using raw connection to peer s> GET /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 405 Method Not Allowed\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Allow: POST\r\n s> Content-Length: 30\r\n s> \r\n s> commands require POST requests Missing Accept header results in 406 $ send << EOF > httprequest POST api/$HTTPV2/ro/customreadonly > user-agent: test > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 406 Not Acceptable\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 85\r\n s> \r\n s> client MUST specify Accept header with value: application/mercurial-exp-framing-0002\n Bad Accept header results in 406 $ send << EOF > httprequest POST api/$HTTPV2/ro/customreadonly > accept: invalid > user-agent: test > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: invalid\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 406 Not Acceptable\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 85\r\n s> \r\n s> client MUST specify Accept header with value: application/mercurial-exp-framing-0002\n Bad Content-Type header results in 415 $ send << EOF > httprequest POST api/$HTTPV2/ro/customreadonly > accept: $MEDIATYPE > user-agent: test > content-type: badmedia > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: badmedia\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 415 Unsupported Media Type\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 88\r\n s> \r\n s> client MUST send Content-Type header with value: application/mercurial-exp-framing-0002\n Request to read-only command works out of the box $ send << EOF > httprequest POST api/$HTTPV2/ro/customreadonly > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name eos customreadonly > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> *\r\n (glob) s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x0e\x00\x00\x01\x00\x01\x01\x11customreadonly s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: application/mercurial-exp-framing-0002\r\n s> Transfer-Encoding: chunked\r\n s> \r\n s> 25\r\n s> \x1d\x00\x00\x01\x00\x02\x01Bcustomreadonly bytes response s> \r\n s> 0\r\n s> \r\n Request to read-write command fails because server is read-only by default GET to read-write request yields 405 $ send << EOF > httprequest GET api/$HTTPV2/rw/customreadonly > user-agent: test > EOF using raw connection to peer s> GET /api/exp-http-v2-0001/rw/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 405 Method Not Allowed\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Allow: POST\r\n s> Content-Length: 30\r\n s> \r\n s> commands require POST requests Even for unknown commands $ send << EOF > httprequest GET api/$HTTPV2/rw/badcommand > user-agent: test > EOF using raw connection to peer s> GET /api/exp-http-v2-0001/rw/badcommand HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 405 Method Not Allowed\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Allow: POST\r\n s> Content-Length: 30\r\n s> \r\n s> commands require POST requests SSL required by default $ send << EOF > httprequest POST api/$HTTPV2/rw/customreadonly > user-agent: test > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/rw/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 403 ssl required\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Length: 17\r\n s> \r\n s> permission denied Restart server to allow non-ssl read-write operations $ killdaemons.py $ cat > server/.hg/hgrc << EOF > [experimental] > web.apiserver = true > web.api.http-v2 = true > [web] > push_ssl = false > allow-push = * > EOF $ hg -R server serve -p $HGPORT -d --pid-file hg.pid -E error.log $ cat hg.pid > $DAEMON_PIDS Authorized request for valid read-write command works $ send << EOF > httprequest POST api/$HTTPV2/rw/customreadonly > user-agent: test > accept: $MEDIATYPE > content-type: $MEDIATYPE > frame 1 1 stream-begin command-name eos customreadonly > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/rw/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> content-length: 22\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x0e\x00\x00\x01\x00\x01\x01\x11customreadonly s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: application/mercurial-exp-framing-0002\r\n s> Transfer-Encoding: chunked\r\n s> \r\n s> 25\r\n s> \x1d\x00\x00\x01\x00\x02\x01Bcustomreadonly bytes response s> \r\n s> 0\r\n s> \r\n Authorized request for unknown command is rejected $ send << EOF > httprequest POST api/$HTTPV2/rw/badcommand > user-agent: test > accept: $MEDIATYPE > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/rw/badcommand HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 404 Not Found\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 42\r\n s> \r\n s> unknown wire protocol command: badcommand\n debugreflect isn't enabled by default $ send << EOF > httprequest POST api/$HTTPV2/ro/debugreflect > user-agent: test > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/debugreflect HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> user-agent: test\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> makefile('rb', None) s> HTTP/1.1 404 Not Found\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 34\r\n s> \r\n s> debugreflect service not available Restart server to get debugreflect endpoint $ killdaemons.py $ cat > server/.hg/hgrc << EOF > [experimental] > web.apiserver = true > web.api.debugreflect = true > web.api.http-v2 = true > [web] > push_ssl = false > allow-push = * > EOF $ hg -R server serve -p $HGPORT -d --pid-file hg.pid -E error.log $ cat hg.pid > $DAEMON_PIDS Command frames can be reflected via debugreflect $ send << EOF > httprequest POST api/$HTTPV2/ro/debugreflect > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name have-args command1 > frame 1 1 0 command-argument 0 \x03\x00\x04\x00fooval1 > frame 1 1 0 command-argument eoa \x04\x00\x03\x00bar1val > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/debugreflect HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> content-length: 54\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x08\x00\x00\x01\x00\x01\x01\x12command1\x0b\x00\x00\x01\x00\x01\x00 \x03\x00\x04\x00fooval1\x0b\x00\x00\x01\x00\x01\x00"\x04\x00\x03\x00bar1val s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 322\r\n s> \r\n s> received: 1 2 1 command1\n s> ["wantframe", {"state": "command-receiving"}]\n s> received: 2 0 1 \x03\x00\x04\x00fooval1\n s> ["wantframe", {"state": "command-receiving"}]\n s> received: 2 2 1 \x04\x00\x03\x00bar1val\n s> ["runcommand", {"args": {"bar1": "val", "foo": "val1"}, "command": "command1", "data": null, "requestid": 1}]\n s> received: <no frame>\n s> {"action": "noop"} Multiple requests to regular command URL are not allowed $ send << EOF > httprequest POST api/$HTTPV2/ro/customreadonly > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name eos customreadonly > frame 3 1 0 command-name eos customreadonly > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/customreadonly HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> content-length: 44\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x0e\x00\x00\x01\x00\x01\x01\x11customreadonly\x0e\x00\x00\x03\x00\x01\x00\x11customreadonly s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 46\r\n s> \r\n s> multiple commands cannot be issued to this URL Multiple requests to "multirequest" URL are allowed $ send << EOF > httprequest POST api/$HTTPV2/ro/multirequest > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name eos customreadonly > frame 3 1 0 command-name eos customreadonly > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/multirequest HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> *\r\n (glob) s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x0e\x00\x00\x01\x00\x01\x01\x11customreadonly\x0e\x00\x00\x03\x00\x01\x00\x11customreadonly s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: application/mercurial-exp-framing-0002\r\n s> Transfer-Encoding: chunked\r\n s> \r\n s> *\r\n (glob) s> \x1d\x00\x00\x01\x00\x02\x01Bcustomreadonly bytes response s> \r\n s> 25\r\n s> \x1d\x00\x00\x03\x00\x02\x01Bcustomreadonly bytes response s> \r\n s> 0\r\n s> \r\n Interleaved requests to "multirequest" are processed $ send << EOF > httprequest POST api/$HTTPV2/ro/multirequest > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name have-args listkeys > frame 3 1 0 command-name have-args listkeys > frame 3 1 0 command-argument eoa \x09\x00\x09\x00namespacebookmarks > frame 1 1 0 command-argument eoa \x09\x00\x0a\x00namespacenamespaces > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/multirequest HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> content-length: 93\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x08\x00\x00\x01\x00\x01\x01\x12listkeys\x08\x00\x00\x03\x00\x01\x00\x12listkeys\x16\x00\x00\x03\x00\x01\x00" \x00 \x00namespacebookmarks\x17\x00\x00\x01\x00\x01\x00" \x00\n s> \x00namespacenamespaces s> makefile('rb', None) s> HTTP/1.1 200 OK\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: application/mercurial-exp-framing-0002\r\n s> Transfer-Encoding: chunked\r\n s> \r\n s> 8\r\n s> \x00\x00\x00\x03\x00\x02\x01B s> \r\n s> 26\r\n s> \x1e\x00\x00\x01\x00\x02\x01Bbookmarks \n s> namespaces \n s> phases s> \r\n s> 0\r\n s> \r\n Restart server to disable read-write access $ killdaemons.py $ cat > server/.hg/hgrc << EOF > [experimental] > web.apiserver = true > web.api.debugreflect = true > web.api.http-v2 = true > [web] > push_ssl = false > EOF $ hg -R server serve -p $HGPORT -d --pid-file hg.pid -E error.log $ cat hg.pid > $DAEMON_PIDS Attempting to run a read-write command via multirequest on read-only URL is not allowed $ send << EOF > httprequest POST api/$HTTPV2/ro/multirequest > accept: $MEDIATYPE > content-type: $MEDIATYPE > user-agent: test > frame 1 1 stream-begin command-name eos unbundle > EOF using raw connection to peer s> POST /api/exp-http-v2-0001/ro/multirequest HTTP/1.1\r\n s> Accept-Encoding: identity\r\n s> accept: application/mercurial-exp-framing-0002\r\n s> content-type: application/mercurial-exp-framing-0002\r\n s> user-agent: test\r\n s> content-length: 16\r\n s> host: $LOCALIP:$HGPORT\r\n (glob) s> \r\n s> \x08\x00\x00\x01\x00\x01\x01\x11unbundle s> makefile('rb', None) s> HTTP/1.1 403 Forbidden\r\n s> Server: testing stub value\r\n s> Date: $HTTP_DATE$\r\n s> Content-Type: text/plain\r\n s> Content-Length: 53\r\n s> \r\n s> insufficient permissions to execute command: unbundle $ cat error.log