https://bugs.exim.org/show_bug.cgi?id=2907
Bug ID: 2907
Summary: TCP bogus retries after TFO
Product: Exim
Version: N/A
Hardware: x86
OS: Linux
Status: NEW
Severity: bug
Priority: medium
Component: Transports
Assignee: unallocated@???
Reporter: jgh146exb@???
CC: exim-dev@???
Multiple people are reporting intermittent delivery failures to Google.
The syndrome is, for a fairly large message, a hang at the end of data
transmission. Exim never gets an SMTP-level ack for the data. This can
be either DATA or BDAT.
A packet capture obtained is implying a Linux kernel bug, either in the TCP
implementation or transmit-offload to the NIC. I'm treating this as a tracking
bug at this time.
Disabling use of TFO (in the smtp transport) seems to be a sufficient
workaround. It happens that Google is the major SMTP peer supporting TFO,
but I would expect the issue to affect any such peer.
--
You are receiving this mail because:
You are on the CC list for the bug.
