[chrony-users] chrony and ntpd xleave interoperability

[FUSTE Emmanuel <emmanuel.fuste@xxxxxxxxxxxxxxx>]

Hello,

First, my apologies for the fingers crossing on chrony-dev when I tried 
to subscribe to chrony-users...

I'm doing some tests to replace ntpd by chrony on some servers groups.
Theses servers use a peer association with interleave option.

When I try to do the same with ntpd on one side and chrony on the other, 
things go bad.
At best, chrony got a working association with interleave status with 
very long response time.
On the ntpd side, the association never work. The chrony server never 
get the "reach" state and the reach counter is stuck a zero.

As soon as I remove  the xleave option on the ntpd side, all start 
immediately to work as expected.

ntpd :
peer y.y.y.y minpoll 5 maxpoll10 xleave
restrict y.y.y.y notrap nomodify noquery

chrony :
peer x.x.x.x xleave minpoll 5  maxpoll 10
allow x.x.x.0/24

Since yesterday, I had removed the xleave option on the ntpd side.
All was good on the two sides.
So I tried to reactivate the xleave option
-> Boom it works !!!

I restarted chrony
-> ntpd logged "revceive: KoD packet from 192.54.145.235 has a zero org 
or rec timestamp. Ignoring."
and four minute later "y.y.y.y 8613 83 unreacheable"
The previously working assoc is now dead.
No working assoc from chrony.

So I restarted ntpd
-> chrony start to see the other server (ntpdata) but never reach a good 
state.
-> ntpd does not reach the "reach" state.

remove the xleave from ntpd and restart
-> all is still stuck
restart chrony
->  ntpd start to see the chrony server, reach state increment, and 
reach a "backup" condition. All is good on the chrony side.

Re-add xleave option on ntpd side.
unreach counter increment, flash=1606 so packet_bogus...
on the chrony side, "Total valid RX" no longer increment...

I'm lost.

chrony 3.2
ntp-4.2.8p8, ntp-4.2.8p10

Could I normally expect xleave interoperability between chrony and ntpd 
or it is something too much "implementation specific" ?

Emmanuel.