This article describes why flowd is crashing and generating core dumps after enabling SSL forward proxy with TLS 1.3.

During SSL decryption process, When TLS session on SSL-T [terminator on the client side] updates the session cache and releases the session-cache resources, proxy session information is getting cleared, if in between TLS 1.3 new ticket is received on SSL-I [initiator on the server side], due to timing synchronization issue flowd process crashes and generates the core dumps. This issue happens when the SSL_I (SRX to Server SSL connection) is using TLS1.3 and SSL_T (Client to SRX SSL connection)is using TLS1.2.

-If the SRX is configured in high-availability, the RG1+ (data-plane) will fail-over to the secondary node.

-If the SRX is configured stand-alone, there will be temporary traffic interruption until the flowd process is restored automatically.

> To check the core files, execute the CLI command:
User@DEVICE> show system core-dump

This issue is being tracked in PR 1788673.

Workaround:

The TLS 1.3 session ticket storing causing the issue. Switch to TLS 1.2 to avoid this issue:
user@device# set services ssl proxy profile <profile-name> protocol-version tls12-and-lower

OR

Disable the session resumption:
user@device# set services ssl proxy profile ssl-proxy actions disable-session-resumption

Resolved in: 

junos:22.4R3-S2

junos:23.2R2

junos:23.4R2

junos:24.1R1