On Mon, Feb 15, 2016 at 4:33 PM, Magnus Hagander <magnus(at)hagander(dot)net> wrote:
>
>
> On Mon, Feb 15, 2016 at 6:56 AM, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
> wrote:
>>
>> On Mon, Feb 15, 2016 at 9:24 AM, Michael Paquier
>> <michael(dot)paquier(at)gmail(dot)com> wrote:
>> > Another possibility to consider would be to mention that in the
>> > documentation on back-branches, and simply block its use on HEAD. For
>> > the quorum sync patch (which is far from baked yet), we are going to
>> > need two extra special characters to identify quorum and priority
>> > groups of nodes, so we would need to provide the same solution as for
>> > commas.
>>
>> Horiguchi-san has just mentioned in [1] that one can use double-quotes
>> in s_s_names around a node name to bypass the problem. That's
>> undocumented though, so wouldn't we want to mention that at least?
>> [1]:
>> http://www.postgresql.org/message-id/20160215.141102.28792569.horiguchi.kyotaro@lab.ntt.co.jp
>>
>
> I think that combined with option 4 (comment the possible caveat) is the
> best way forward. Especially since we have this ability, there is no need to
> restrict it and possibly break existing applications. But documenting it
> would clearly be an advantage.
OK, I just typed up a doc patch that I think follows those lines, here
is the paragraph:
+ When a standby node includes in its <varname>application_name</>
+ value a comma (<literal>,</>), it can not be chosen as a synchronous
+ standby because <varname>synchronous_standby_name</> uses as a
+ separator character a comma. This can be avoided by double-quoting
+ the standby name in the list provided by this parameter. The use
+ of commas in <varname>application_name</> is not recommended to avoid
+ any confusion though.
--
Michael