Remove sub from rsubs sublist when user UNSUBs.
Fix bench test that was not actually creating a SUB per request
in the Benchmark_Gateways_Requests_CreateOneSubForEach test.
Also UNSUBs older SUBs after a certain threshold to simulate
actual req/reply.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
This addresses the following race:
- client connection creates a subscription on a reply subject
- client connection sends a request
- server sends the subscription to inbound gateway
- server sends the message to outbound gateway (those may be
to different servers)
- receiving server sends to sub interested in request subject
- app sends reply
- its server then check for interest on the reply's subject
In interestOnly mode, there is a possibility that this server
has not received the interest on the reply subject yet and would
then drop the reply.
This PR detects above scenario and will prefix the reply subject
to identify the origin cluster if it is detected that the last
subscription from the sending connection was created less than
a second ago.
Once the destination has this prefix, the destination cluster
will always send back that message to origin cluster even if
there is no registered interest.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
A server that creates a LeafNode connection to a remote cluster
will now be notified of all possible LeafNode URLs in that cluster.
The list is updated when nodes in the cluster come and go.
Also support for advertise address, similar to cluster, gateway, etc..
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
What is not completed:
1. TLS
2. config to bind local account.
3. Info updates for solicitor to track topology changes like a client.
4. CONNECT sent after INFO for nonce authroization.
5. Authorization
6. Services and Streams tests.
7. config file parsing.
Signed-off-by: Derek Collison <derek@nats.io>
Reworked fan in logic. We do not hold locks during IO, either read or write.
On scenarios where we can get behind mostly due to fan-in from fast producers
we detect and create a stall channel. Once we catch up we close the stall channel
to release all blocked producers. Producers have an upper bound on how long
they will be stalled.
Signed-off-by: Derek Collison <derek@nats.io>
If each server has a long list of subscriptions, when the route
is established, sending this list could result in each server
treating the peer as a slow consumer, resulting in a reconnect,
etc..
Also bumping the fan-in threshold for route connections.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
1. Change outbound client structure to be smaller and more cache friendly.
2. Snapshot MaxControlLine into client structure (mcl) to avoid server opts lookup.
Signed-off-by: Derek Collison <derek@nats.io>
There is a possibility that a partial write results in data
not being sent in a timely fashion to a subscription.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
Under some scenarios a client may hit the slow consumer
write deadline during the connecting stage, meaning that
it is possible to consider unhealthy clients could not
finish the TLS handshake as slow consumers.
With this change, we only consider as slow consumers
clients that did connect to the cluster without issues.
Signed-off-by: Waldemar Quevedo <wally@synadia.com>
Converted to authorization error events on different subject.
Add cluster name if gateways are configured and pass in INFO to clients.
Signed-off-by: Derek Collison <derek@nats.io>
Do the swapping to outbound connection only on send.
It means that those subs are stored in the inbound connection and
those are the only type of subs stored there. So on connection close
it is easy to clean them up.
Also instead of having processMsgResults have to return this sub,
simply check the size of r.psubs and if 1, the type of client
associated with it. If gateway, we know we have to do the direct
send.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
This adds a RemoteAddress() method to the ClientAuthentication
interface which would facilitate CustomClientAuthentication
mechanisms considering the remote IP address of the client in
their decisions to allow clients or not
Resolves#832
Signed-off-by: R.I.Pienaar <rip@devco.net>
This is not complete solution and is a bit hacky but is a start
to be able to have service import work at least in some basic
cases.
Also fixed a bug where replySub would not be removed from
connection's list of subs after delivery.
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
Changed account lookup and validation failures to be more understandable by users.
Changed limits to be -1 for unlimited to match jwt pkg.
The limits changed exposed problems with options holding real objects causing issues with reload tests under race mode.
Longer term this code should be reworked such that options only hold config data, not real structs, etc.
Signed-off-by: Derek Collison <derek@nats.io>
When a TLS timeout occurs, it was possible that code trying to
send data would get a timeout error and report it as a Slow Consumer
with the write deadline exceeded error.
Checking that if the connection is TLS and handshake not complete,
it is likely a TLS timeout error instead.
Resolves#835
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
Added update to parse and load operator JWTs.
Changed to add in signing keys from operator JWT to list of trusted keys.
Added URL account resolver.
Added account claim updates by system messages.
Signed-off-by: Derek Collison <derek@nats.io>
Specifically this is to support distributed tracking of number of account connections across clusters.
Gateways may not work yet based on attempts to only generate payloads when we know there is outside interest.
Signed-off-by: Derek Collison <derek@nats.io>
- If/when splitting buffer to pass to queueOutbound(), it has to
be include full protocol.
- Fix counting of total queue subs
- Fix tests
- Send RS- if no plain sub interest even if there is queue sub
interest.
- Removed a one-liner function
Signed-off-by: Ivan Kozlovic <ivan@synadia.com>
