Calls to `filteredStateLocked` can mutate the per-subject state by
recalculating the first sequences. Doing so under a read-only lock can
lead to race conditions, so switch the call sites out to write locks
instead.
Signed-off-by: Neil Twigg <neil@nats.io>
The `streamSnapshot` function was not holding the filestore lock while
resetting and using the hasher. This created a data race and also
resulted in flakes in `TestFileStoreSnapshot`.
Signed-off-by: Neil Twigg <neil@nats.io>
- Use Go 1.21 in nightlies
- Both rand.Seed and rand.Read are both deprecated, remove its use to
fix staticcheck errors
```go
server/client.go:95:2: SA1019: rand.Seed has been deprecated since Go 1.20 and an alternative
has been available since Go 1.0: As of Go 1.20 there is no reason to call Seed with a random value.
Programs that call Seed with a known value to get a specific sequence of results should use
New(NewSource(seed)) to obtain a local random generator. (staticcheck)
server/jetstream_test.go:20399:2: SA1019: rand.Read has been deprecated since Go 1.20
because it shouldn't be used: For almost all use cases, crypto/rand.Read is more appropriate. (staticcheck)
```
This test flakes quite often on the final consumer info check because of
one consumer reporting `PushBound` and the other not. By clearing both,
this should clear up the disagreement.
Signed-off-by: Neil Twigg <neil@nats.io>
In +Go 1.20, the x509.HostnameError changed to be wrapped in a
tls.CertificateVerificationError so sometimes the name would not
be reset.
This would make the `TestGatewayTLSMixedIPAndDNS` test flake sometimes.
Signed-off-by: Waldemar Quevedo <wally@nats.io>
- [x] Branch rebased on top of current main (`git pull --rebase origin
dev`)
- [x] Changes squashed to a single commit (described
[here](http://gitready.com/advanced/2009/02/10/squashing-commits-with-rebase.html))
- [ ] Build is green in Travis CI
### Changes proposed in this pull request:
Removed `qos2` reference from the MQTT outgoing stream name and
subjects.
Rationale: As I was reviewing the documentation and thinking how to move
forward with MQTT, I am inclined to use the `$MQTT_out` stream and
session-specific consumers (as opposed to subscription-specific) for all
QOS-related packet delivery and PI persistence. I think I will be able
to repurpose it "as is", so proposing the generic name change.
Since these (stream, consumer, subscriptions) are newly introduced in
https://github.com/nats-io/nats-server/pull/4349 and have not yet been
released, this is a safe change.
This uses `TempDir` in more tests. Some were trying to manage the
directory manually, others were just not managing a directory at all.
There may be more that need doing.
Signed-off-by: Neil Twigg <neil@nats.io>
- [ ] Link to issue, e.g. `Resolves #NNN`
- [ ] Documentation added (if applicable)
- [x] Tests added
- [x] Branch rebased on top of current main (`git pull --rebase origin
main`)
- [ ] Changes squashed to a single commit (described
[here](http://gitready.com/advanced/2009/02/10/squashing-commits-with-rebase.html))
- [ ] Build is green in Travis CI
- [x] You have certified that the contribution is your original work and
that you license the work to the project under the [Apache 2
license](https://github.com/nats-io/nats-server/blob/main/LICENSE)
Related: https://github.com/nats-io/nats-server/pull/4399
### Changes proposed in this pull request:
- Avoid binding to hardcoded port, as it may conflict if already in use
- Drop TLS from Core Request/Response benchmark, was not showing much
difference, not enough data going through to become a dominant cost
- Add a different core benchmark that exercises TLS by pushing through
lots of data
