On Mon, 24 Feb 2020 at 08:46, Paul Wouters <[email protected]> wrote: > > On Mon, 24 Feb 2020, Antony Antony wrote: > > > I have an idea to sanitize the transient lines such as > > "retransmission; will wait" during testrun. > > > > > > Some test where we need to the retransmission add a special marker e.g > > "ipsec auto --up <conn> #retransmits" I suggest "#retransmits" as a > > convention. and open to short catchy word. It can also be just # > > > > "ipsec auto --up <conn>" #retransmits > > > > I added #retransmits to 14 to 10 tests. > > What is the issue with using slow retransmits, that is now used in many > tests?
Sometimes it is expected to die? It's also possible to force specific behaviour on the first retransmit (altough with IKEv1<->IKEv2 dead, we may need one lesss): timeout-on-retransmit: causes pluto to 'retry' (switch protocol) on the first retransmit delete-on-retransmit: causes pluto to fail on the first retransmit suppress-retransmits: causes pluto to never send retransmits (wait the full timeout) > paul@bofh:~$ grep "slow retrans" libreswan/testing/pluto//*/*.conf| wc -l > 171 > > Paul > _______________________________________________ > Swan-dev mailing list > [email protected] > https://lists.libreswan.org/mailman/listinfo/swan-dev _______________________________________________ Swan-dev mailing list [email protected] https://lists.libreswan.org/mailman/listinfo/swan-dev
