dovecot - May 2018 - Upgrading dovecot 2.2 to 2.3 without downtime when using proxy/director?

Hi,

I have a setup with several dovecot-servers (2.2.35) and I use dovecot
proxy. I upgraded one server to 2.3.1 and got the configs fixed so far
that it started again. But when I tried to add it into the proxying
again with "doveadm director add" I see the following in the logfiles:

May  6 07:19:30 host dovecot: director: Warning: Director
xxx.xxx.xxx.176:9090/out disconnected us with reason: Invalid input: OPTIONS
May  6 07:19:30 $host dovecot: director: Connecting to
xxx.xxx.xxx.171:9090 (as xxx.xxx.xxx.176): Reconnecting after error

xxx.xxx.xxx.176 is the upgraded host, xxx.xxx.xxx.171 is one of the
other hosts (the applicable hosts are in the range xxx.xxx.xxx.171-176).

Do I have to remove all the hosts from the proxy, upgrade them and then
add them again? Or is there a way to handle the upgrade without a
downtime? After all one of the reasons I use several hosts is that I do
not want to have a downtime.

Niels

> On 6 May 2018, at 8.46, Niels Kobsch?tzki <niels at kobschaetzki.net>
wrote:
> 
> Hi,
> 
> I have a setup with several dovecot-servers (2.2.35) and I use dovecot
> proxy. I upgraded one server to 2.3.1 and got the configs fixed so far
> that it started again. But when I tried to add it into the proxying
> again with "doveadm director add" I see the following in the
logfiles:
> 
> May  6 07:19:30 host dovecot: director: Warning: Director
> xxx.xxx.xxx.176:9090/out disconnected us with reason: Invalid input:
OPTIONS
> May  6 07:19:30 $host dovecot: director: Connecting to
> xxx.xxx.xxx.171:9090 (as xxx.xxx.xxx.176): Reconnecting after error
> 
> xxx.xxx.xxx.176 is the upgraded host, xxx.xxx.xxx.171 is one of the
> other hosts (the applicable hosts are in the range xxx.xxx.xxx.171-176).
> 
> Do I have to remove all the hosts from the proxy, upgrade them and then
> add them again? Or is there a way to handle the upgrade without a
> downtime? After all one of the reasons I use several hosts is that I do
> not want to have a downtime.
If you look at .176's error log, do you see an error about
"director_consistent_hashing settings differ between directors"? Have
you set director_consistent_hashing=yes in the old directors? That is needed
now, because the old non-consistent-hashing method is obsoleted. Unfortunately
there's no easy way to upgrade directors to use the consistent hashing
method without stopping the entire ring. The hard way would be to build a
secondary director ring and start moving users to that ring in proxies.

On 15 May 2018, at 12.06, Timo Sirainen <tss at iki.fi>
wrote:
> 
> If you look at .176's error log, do you see an error about
"director_consistent_hashing settings differ between directors"? Have
you set director_consistent_hashing=yes in the old directors? That is needed
now, because the old non-consistent-hashing method is obsoleted. Unfortunately
there's no easy way to upgrade directors to use the consistent hashing
method without stopping the entire ring. The hard way would be to build a
secondary director ring and start moving users to that ring in proxies.
Added https://wiki2.dovecot.org/PasswordDatabase/ExtraFields/Proxy#moving
<https://wiki2.dovecot.org/PasswordDatabase/ExtraFields/Proxy#moving> to
explain further the hard way.

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://dovecot.org/pipermail/dovecot/attachments/20180515/e8df0172/attachment.html>