Github user mtaylor commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/690#discussion_r73333507
--- Diff:
artemis-protocols/artemis-amqp-protocol/src/main/java/org/apache/activemq/artemis/core/protocol/proton/plug/ProtonSessionIntegrationCallback.java
---
@@ -351,18 +353,35 @@ public void serverSend(final Receiver receiver,
recoverContext();
PagingStore store =
manager.getServer().getPagingManager().getPageStore(message.getAddress());
- if (store.isFull() && store.getAddressFullMessagePolicy() ==
AddressFullMessagePolicy.BLOCK) {
- ErrorCondition ec = new
ErrorCondition(AmqpError.RESOURCE_LIMIT_EXCEEDED, "Address is full: " +
message.getAddress());
- Rejected rejected = new Rejected();
- rejected.setError(ec);
- delivery.disposition(rejected);
- connection.flush();
+ if (store.isRejectingMessages()) {
+ // We drop pre-settled messages (and abort any associated Tx)
+ if (delivery.remotelySettled()) {
+ if (serverSession.getCurrentTransaction() != null) {
+ rollbackCurrentTX(false);
+
+ // Spec requires closure of coordinator link on rejection
of presettled messages within a tx
--- End diff --
My opinion is that, this fix is spec compliant and fixes the intended
issue. I'd prefer to merge this and close out the hard/soft limit issue. If
people prefer us to use option 2. then we can create another issue to cover the
switch in policy, any client that does not handle this scenario is surely not
spec compliant?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
