[ARTEMIS-1796] Flush during handshake of AMQP can cause issues with SASL clients - ASF JIRA

XML

Word

Printable

JSON

Type: Bug
Status: Closed
Priority: Major
Resolution: Fixed
Affects Version/s: 2.5.0
Fix Version/s: 2.6.0
Component/s: None
Labels:
None
Environment:

Hide

I found this by setting up a debug point on the handshake. But as I saw it the flush could still happen shortly after causing intermittent issues with SASL clients.

The flush should be deferred until the initial handshake is finished. This is as simple as only doing the flush if the first packet was received, and that would be enough for this.

The solution I'm implementing is to differentiate flushes from scheduled flushes, and doing a regular flush if the first packet was received.

Show
I found this by setting up a debug point on the handshake. But as I saw it the flush could still happen shortly after causing intermittent issues with SASL clients. The flush should be deferred until the initial handshake is finished. This is as simple as only doing the flush if the first packet was received, and that would be enough for this. The solution I'm implementing is to differentiate flushes from scheduled flushes, and doing a regular flush if the first packet was received.

links to

GitHub Pull Request #2002