[ https://issues.apache.org/jira/browse/QPID-8648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17810245#comment-17810245 ]
ASF GitHub Bot commented on QPID-8648: -------------------------------------- dakirily opened a new pull request, #234: URL: https://github.com/apache/qpid-broker-j/pull/234 This PR addresses JIRA [QPID-8648](https://issues.apache.org/jira/browse/QPID-8648), increasing maximal frame size before Open frame during SASL negotiation > [Broker-J] Allow for max frame size >4096 before Open frame (SASL) > ------------------------------------------------------------------ > > Key: QPID-8648 > URL: https://issues.apache.org/jira/browse/QPID-8648 > Project: Qpid > Issue Type: Improvement > Components: Broker-J > Affects Versions: qpid-java-broker-9.0.0 > Reporter: Dan Langford > Priority: Major > Fix For: qpid-java-broker-9.1.1 > > > some modern authentication options (XOAUTH2 + JWT) require frames larger then > 4096. consider if the max frame size (before an Open frame negotiation) > should be larger or should be configurable with some sort of configuration or > env variable. > > from a discussion on the mailing list > {quote}The SASL process occurs first, before the Open frame. The Open frames > are what carries each peers advertised max frame size, mainly aimed at > later message deliveries. The AMQP 1.0 spec defines before this > however that the SASL frames can be at-most the 'min max frame size', > which is fixed at 512 bytes, with no way to negotiate anything larger. > As you can probably tell, that presents a problem if things in the > SASL negotiation want to be larger, such as is likely in e.g a newer > XOAUTH2 mechanism that didnt exist when that decision was originally > made. > To simply allow some of these newer alternative mechs to work, it was > decided to just allow things to exceed the 512byte limit since both > sides would have to already agree on using a given mech to begin with, > so doing an alternative like creating a custom multi-challenge > batching sequence to shuffle the bytes wasnt really going to be adding > much except significant complexity. > It appears broker-j allows up to 4096, and you have now found > something to exceed even that. It doesnt look like it allows > configuring it, but increasing that seems to be the only option that > would help here. > {quote} -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org