When moving streams, we could check too soon and be in a gap where the
replica peer has not registered a catchup request but had made contact
via the NRG layer.
This would cause us to think the replica was caught up, incorrectly, and
drop our leadership, which would cancel any catchup requests.
Signed-off-by: Derek Collison <derek@nats.io>
This would cause us to think the replica was caughtup incorrectly and drop our leadership, which would cancel any cacthup requests.
Signed-off-by: Derek Collison <derek@nats.io>
Fix for a bug that would allow old leaders of pull based durables to
delete a consumer from an inactivity threshold timer inadvertently.
Signed-off-by: Derek Collison <derek@nats.io>
We violated the locking pattern, so we now make sure we do this in a
separate Go routine and put checks to only run it once.
Signed-off-by: Derek Collison <derek@nats.io>
If we send an event when entering lame duck mode, other nodes will mark
the server as offline immediately, therefore R1 assets will not be
placed onto that node. This is not a problem with R3 or higher because
an LDM server operates as a Raft observer only and therefore cannot take
the leadership role from an election, but R1 assets can in theory be
placed onto any node that is not marked as offline.
A final shutdown event will still be sent when the server actually shuts
down so there is no change there.
Signed-off-by: Neil Twigg <neil@nats.io>
We violated the locking pattern, so we now make sure we do this in a separate Go routine and put checks to only run it once.
Signed-off-by: Derek Collison <derek@nats.io>
This should hopefully catch some consumer scaling situations more
reliably, including cases where the consumer filter subjects no longer
match those of the stream after being scaled down to R1 or after a
cluster restart. I've also added a test to test whether filtered
consumers will scale properly even when the stream subject orphans them.
Signed-off-by: Neil Twigg <neil@nats.io>
Allows again empty configs and improves support for detecting some types
of invalid configs:
- Adds reporting the line with the bad key position that makes the
config invalid.
```
nats-server: config is invalid (foo.conf:1:2)
nats-server: error parsing include file 'included.conf', config is invalid (included.conf:2:2)
```
- Fixes a few tests with trailing braces which were being handled as
keys and ignored before.
- Adds reporting the line with the bad key position
that makes the config invalid.
- Fixes a few tests with trailing braces which were
being handled as keys and ignored before.
Signed-off-by: Waldemar Quevedo <wally@nats.io>
If the failed state of clfs drifts between leaders and followers,
replicas could discard and skip messages possibly incorrectly. This will
force sync if we have a non-zero clfs state when a leader takes over.
Signed-off-by: Derek Collison <derek@nats.io>
This PR backports the OCSP Peer feature option (as in 2.10 train) and
includes two fixes for the existing OCSP Staple feature.
OCSP Staple:
1. Fixed and clarified how NATS Server determines its own Issuer CA when
obtaining and validating an OCSP Response for subsequent staple
2. Eliminated problematic assumption that all node peers are issued by
same CA when NATS Server validates ROUTE and GATEWAY peer nodes
3. Added OCSP Response effectivity checks on ROUTE and GATEWAY
peer-presented staple
Note for #3: Allowed host clock skew between node peers set at
30-seconds. If the OCSP Response contains an empty assertion for
NextUpdate, NATS Server will default to 1-hour validity (after
ThisUpdate). It is recommended that CA OCSP Responder should assert
NextUpdate.
On reload some of the imports from the system account where going
missing on reload, this adds them back after a reload:
```
$SYS.REQ.SERVER.PING.CONNZ
$SYS.REQ.ACCOUNT.PING.STATZ
$SYS.REQ.ACCOUNT.PING.CONNZ
```
This makes configuration files that are empty, or read and processed by
the parser but with no detected values now return an error.
Fixes#4343
Backport from dev branch
(https://github.com/nats-io/nats-server/pull/4347)
This makes configuration files that are empty, or read and processed
by the parser but with no detected values now return an error.
Signed-off-by: Waldemar Quevedo <wally@nats.io>
Do not hold onto no interest subjects from a client in the unlocked cache.
If sending lots of different subjects all with no interest performance could be affected.
Signed-off-by: Derek Collison <derek@nats.io>
Resolves#4341
Three issues were found and resolved.
1. Some purge replays after recovery could execute full purge.
2. Callback was registered without lock, which could lead to skew.
3. Cluster reset could stop stream store and recreate it, which could
lead to double accounting.
Signed-off-by: Derek Collison <derek@nats.io>
