Limit the number WebSocket events queued from external clients to
dispatcher instances to produce back pressure to the clients if
Zeek's IO loop is overloaded.
Explicitly notify the internal thread about the shutdown via the
inproc socket pair. This ensures that the internal thread processes
all previous messages on the inproc socket before terminating.
This fixes the scenario where a backend is created, a few messages published
and then immediately terminated as can be done with WebSocket clients.
Previously, some of the messages published might have still been in the
inproc socket's queue and were simply discarded.
Adds the same test for Broker and ZeroMQ backends.
This adds a cluster backend implementation using broker's hub primitive
to connect WebSocket clients with the local broker endpoint for pub/sub
functionality.
If the server doesn't include a responseToken within negTokenResp,
assume there won't be signing or sealing happening on the
connection. Don't switch into MS_KRB5 mode.
Closes#4275
* origin/topic/awelzel/generic-metadata-pre-work:
cluster/Backend: Add name and lookup component tag
cluster/Event: Hide members behind accessors
cluster/PublishEvent:: Make event non-const
broker/Manager: Re-use broker serializer for conversion
EventMgr: Add Dispatch() with handler and args
plugin/Manager: Fix MetaHookPre and MetaHookPost using HOOK_CALL_FUNCTION
For scheduled events, the event timestamp is the intended timestamp. If
we force timer expiration, the timestamp might be in the future. Today,
this happens on shutdown. This change guarantees that event timestamps
are never set beyond network time.
Allow users to call event_mgr.Dispatch(handler, args) instead of
constructing the Event instance themselves. Deprecate the old API
and replace users.
There's a subtle change that net_done() may be propagated via
auto_publish() now, but that still needs opt-in from script land
and likely no one did that, or else they'd expected to have it
work anyhow.
This test is very sensitive to new enum values that perturb the final
order of the output. Listing the commits that modify that file in the
past clearly shows changes to the baseline happen when a new component
with an enum value is introduced.
40f04d4ccf broker: Add WebSocketShim backend
51836d08ae protocol: Add StreamEvent analyzer
fdde1e9841 cluster/serializer: Add binary-serialization-format
6fb73aa9da broker: Add shim plugin adding a backend component
85ca59484b postgresql: Initial parser implementation
efc2681152 WebSocket: Introduce new analyzer and log
...
That's just distraction. Point fix the test to sort the analyzers before
writing the final log line.
I've previously attempted this in a more complete fashion in PR #3365
(issue #3364), but seems this direct fix might just make more sense.
This test has become flaky after #4295. It hasn't been failing regularly
enough locally to make it approachable. Annotate with
BTEST_BG_RUN_SLEEP=1 to revert to pre #4295 behavior, hopefully
fixing the flakiness as a side-effect.
* origin/topic/timw/4281-unconditionally-forward-out-of-snap:
Make SNAP analyzer use both OUI and protocol for forwarding
Change packet analyzer identifiers to be 64-bit
The logic was template-based so far, which wasn't great because: (1)
conceptually, it models the Spicy types at the wrong layer (C++ rather
than HILTI types), and (2) stopped working with some recent Spicy
updates (which we have temporarily reverted in the meantime to keep
Zeek working).
The new code is based on HILTI's runtime type information and the
corresponding introspection API, pretty much like `spicy-dump` works
as well. This is the recommended approach for working with HILTI
values, and generally much cleaner.
This is on top of https://github.com/zeek/zeek/pull/4300.
Calling collect_metrics() from a script would not invoke metric
callbacks, resulting in most of the process metrics to be zero
when a Zeek process isn't scraped via Prometheus.
Fixes#4309
Due to prefix matching, worker-1's node_topic() also matched worker-10,
worker-11, etc. Suffix the node topic with a `.`. The original implementation
came from NATS, where subjects are separated by `.`.
Adapt nodeid_topic() for consistency.
These test were very sensible to the speed at which ZeroMQ distributes
subscriptions in the cluster and showed to be unreliably when testing with
zeek/btest#113.
The main fix here is to have individual WebSocket clients subscribe to unique
topics, e.g /test/client-0 and /test/client-1, instead of just a shared topic.
This ensures the WebSocket handshake completes only when they observed their
own subscriptions and not prematurely when observing the shared topic.
This seems mainly relevant for tests: In the real world one shouldn't
rely on subscription visibility - you miss messages if you're too late
to the party.
When two workers connect to zeek.cluster.worker, the central ZeroMQ
proxy would not propagate unsubscription information to other nodes
once they both left. Set ZMQ_XPUB_VERBOSER on the proxies XPUB socket
for visibility.
This btest uses the exit() BIF to shut down, which immediately calls
::exit() and kills Zeek without doing any shutdown. This will sometimes
leave the thread running the storage manager, which causes TSan to
complain about a thread leak. Switch to use the terminate() BIF instead
which cleanly shuts down all of Zeek.
Bump to the latest btest master version and set BTEST_BG_RUN_SLEEP
environment variable and to 0. This makes btest-bg-run return
immediately instead of delaying by 1 second.
