Commit Graph

24 Commits

Author SHA1 Message Date
Harshavardhana ff80cfd83d
move Make,Delete,Head,Heal bucket calls to websockets (#18951) 2024-02-02 14:54:54 -08:00
Klaus Post ce0cb913bc
Fix ineffective recycling (#18952)
Recycle would always be called on the dummy value `any(newRT())` instead of the actual value given to the recycle function.

Caught by race tests, but mostly harmless, except for reduced perf.

Other minor cleanups. Introduced in #18940 (unreleased)
2024-02-02 08:48:12 -08:00
Klaus Post b192bc348c
Improve object reuse for grid messages (#18940)
Allow internal types to support a `Recycler` interface, which will allow for sharing of common types across handlers.

This means that all `grid.MSS` (and similar) objects are shared across in a common pool instead of a per-handler pool.

Add internal request reuse of internal types. Add for safe (pointerless) types explicitly.

Only log params for internal types. Doing Sprint(obj) is just a bit too messy.
2024-02-01 12:41:20 -08:00
Harshavardhana 6440d0fbf3
move a collection of peer APIs to websockets (#18936) 2024-02-01 10:47:20 -08:00
Klaus Post 6da4a9c7bb
Improve tracing & notification scalability (#18903)
* Perform JSON encoding on remote machines and only forward byte slices.
* Migrate tracing & notification to WebSockets.
2024-01-30 12:49:02 -08:00
Harshavardhana 9987ff570b avoid calling close for nil inbound/outblock channels 2024-01-28 19:56:32 -08:00
Harshavardhana 7743d952dc
fix: incomingBytes() to update via handleMessages() (#18891)
previous change #18880 was incomplete
2024-01-28 14:35:53 -08:00
Harshavardhana 1d3bd02089
avoid close 'nil' panics if any (#18890)
brings a generic implementation that
prints a stack trace for 'nil' channel
closes(), if not safely closes it.
2024-01-28 10:04:17 -08:00
Klaus Post 38de8e6936
grid: Simpler reconnect logic (#18889)
Do not rely on `connChange` to do reconnects.

Instead, you can block while the connection is running and reconnect 
when handleMessages returns.

Add fully async monitoring instead of monitoring on the main goroutine 
and keep this to avoid full network lockup.
2024-01-28 08:46:15 -08:00
Harshavardhana c51f9ef940
fix: regression in internode bytes counting (#18880)
wire up missing metrics since #18461

Bonus: fix trace output inconsistency
2024-01-27 00:25:49 -08:00
Harshavardhana 74851834c0
further bootstrap/startup optimization for reading 'format.json' (#18868)
- Move RenameFile to websockets
- Move ReadAll that is primarily is used
  for reading 'format.json' to to websockets
- Optimize DiskInfo calls, and provide a way
  to make a NoOp DiskInfo call.
2024-01-25 12:45:46 -08:00
Harshavardhana e377bb949a
migrate bootstrap logic directly to websockets (#18855)
improve performance for startup sequences by 2x for 300+ nodes.
2024-01-24 13:36:44 -08:00
Klaus Post 6968f7237a
Add separate grid reconnection mutex (#18862)
Add separate reconnection mutex

Give more safety around reconnects and make sure a state change isn't missed.

Tested with several runs of `λ go test -race -v -count=500`

Adds separate mutex and doesn't mix in the testing mutex.
2024-01-24 11:49:39 -08:00
Klaus Post feeeef71f1
Add extra protection for grid reconnects (#18840)
Race checks would occasionally show race on handleMsgWg WaitGroup by debug messages (used in test only).

Use the `connMu` mutex to protect this against concurrent Wait/Add.

Fixes #18827
2024-01-22 09:39:06 -08:00
Klaus Post 83bf15a703
grid: Return rejection reason (#18834)
When rejecting incoming grid requests fill out the rejection reason and log it once.

This will give more context when startup is failing. Already logged after a retry on caller.
2024-01-19 10:35:24 -08:00
Harshavardhana dd2542e96c
add codespell action (#18818)
Original work here, #18474,  refixed and updated.
2024-01-17 23:03:17 -08:00
Klaus Post 479940b7d0
Deallocate huge read buffers (#18813)
If a message buffer is excessively huge, release it back so it isn't kept around forever.
2024-01-17 11:47:42 -08:00
Harshavardhana da55499db0
fix: reject clients that do not send proper payload (#18701) 2023-12-22 01:26:17 -08:00
Klaus Post 8fc200c0cc
Truncate long traces for internode communication (#18593)
Prevent excessively long request traces.
2023-12-05 12:16:48 -08:00
Klaus Post 860fc200b0
Local and Remote hosts swapped in grid traces (#18574)
Local and Remote hosts swapped in grid trace

A bit counter-intuitive, but simple fix.
2023-12-01 08:04:08 -08:00
Klaus Post 5f971fea6e
Fix Mux Connect Error (#18567)
`OpMuxConnectError` was not handled correctly.

Remove local checks for single request handlers so they can 
run before being registered locally.

Bonus: Only log IAM bootstrap on startup.
2023-12-01 00:18:04 -08:00
Klaus Post 0bb81f2e9c
Always remove subroute when queuing message on the connection. (#18550) 2023-11-28 11:22:29 -08:00
Klaus Post ca488cce87
Add detailed parameter tracing + custom prefix (#18518)
* Allow per handler custom prefix.
* Add automatic parameter extraction
2023-11-26 01:32:59 -08:00
Klaus Post 51aa59a737
perf: websocket grid connectivity for all internode communication (#18461)
This PR adds a WebSocket grid feature that allows servers to communicate via 
a single two-way connection.

There are two request types:

* Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small
  roundtrips with small payloads.

* Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`,
  which allows for different combinations of full two-way streams with an initial payload.

Only a single stream is created between two machines - and there is, as such, no
server/client relation since both sides can initiate and handle requests. Which server
initiates the request is decided deterministically on the server names.

Requests are made through a mux client and server, which handles message
passing, congestion, cancelation, timeouts, etc.

If a connection is lost, all requests are canceled, and the calling server will try
to reconnect. Registered handlers can operate directly on byte 
slices or use a higher-level generics abstraction.

There is no versioning of handlers/clients, and incompatible changes should
be handled by adding new handlers.

The request path can be changed to a new one for any protocol changes.

First, all servers create a "Manager." The manager must know its address 
as well as all remote addresses. This will manage all connections.
To get a connection to any remote, ask the manager to provide it given
the remote address using.

```
func (m *Manager) Connection(host string) *Connection
```

All serverside handlers must also be registered on the manager. This will
make sure that all incoming requests are served. The number of in-flight 
requests and responses must also be given for streaming requests.

The "Connection" returned manages the mux-clients. Requests issued
to the connection will be sent to the remote.

* `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)`
   performs a single request and returns the result. Any deadline provided on the request is
   forwarded to the server, and canceling the context will make the function return at once.

* `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)`
   will initiate a remote call and send the initial payload.

```Go
// A Stream is a two-way stream.
// All responses *must* be read by the caller.
// If the call is canceled through the context,
//The appropriate error will be returned.
type Stream struct {
	// Responses from the remote server.
	// Channel will be closed after an error or when the remote closes.
	// All responses *must* be read by the caller until either an error is returned or the channel is closed.
	// Canceling the context will cause the context cancellation error to be returned.
	Responses <-chan Response

	// Requests sent to the server.
	// If the handler is defined with 0 incoming capacity this will be nil.
	// Channel *must* be closed to signal the end of the stream.
	// If the request context is canceled, the stream will no longer process requests.
	Requests chan<- []byte
}

type Response struct {
	Msg []byte
	Err error
}
```

There are generic versions of the server/client handlers that allow the use of type
safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 17:09:35 -08:00