Re: Patch: Implement failover on libpq connect level.

On Fri, Sep 9, 2016 at 4:49 AM, Victor Wagner <vitus@wagner.pp.ru> wrote:
> Random permutation is much more computationally complex than random
> picking.

It really isn't.  The pseudocode given at
https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle is all of 4
lines long, and one of those lines is a comment.  The C implementation
is longer, but not by much.

Mind you, I haven't read the patch, so I don't know whether using
something like this would make the code simpler or more complicated.
However, if the two modes of operation are (1) try all hosts in random
order and (2) try all hosts in the listed order, it's worth noting
that the code for the second thing can be used to implement the first
thing just by shuffling the list before you begin.

> So, using random permutation instead  of random pick wouln't help.
> And keeping somewhere number of elements in the list wouldn't help
> either unless we change linked list to completely different data
> structure which allows random access.

Is there a good reason not to use an array rather than a linked list?

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company