Connections stall, fail, or take a long time to complete due to excessive TCP retransmissions

Solution Verified - Updated -

Issue

  • When using iptables, a lot of retransmissions are performed, the connection is reset.
  • When using firewalld, connections are stalled (hung) for long periods of time and may be reset.
  • TCP packets/segments containing select acknowledgement (SACK) options are ignored when iptables is being used
  • System is not sending TCP Fast Retransmissions/Retransmits after receiving three Duplicate Acknowledgements when iptables is used
  • SCP transfers are intermittently failing
  • TCP transfer stalls and eventually disconnects with "broken pipe" (EPIPE) message
  • In Wireshark/tshark, SACK Left Edge (SRE) and SACK Right Edge (SRE) on Dup-ACK packets are much larger than the relative SEQ and ACK numbers

Environment

  • Red Hat Enterprise Linux
  • TCP (Transmission Control Protocol)
  • SACK (Selective Acknowledgements)
  • iptables or firewalld firewall

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase of over 48,000 articles and solutions.

Current Customers and Partners

Log in for full access

Log In