[python-ldap] reconnect management

[Michael Ströder]

Florian Friesdorf wrote:
> On Fri, 22 Apr 2011 17:00:12 +0200, Michael Ströder <michael at stroeder.com> wrote:
>> (I split your posting to several postings with distinct mail subjects)
>>
>> Florian Friesdorf wrote:
>>> I am interest in more
>>> sophisticated reconnect management, which was also mentioned recently on
>>> the mailing list. 
>>
>> Can you please point me to your former posting. I may have missed it.
>>
>> What exactly do you want to improve beyond ReconnectLDAPObject?
> 
> It was a post by Bob Brandt, you answered here:
> http://mail.python.org/pipermail/python-ldap/2011q2/002850.html
> 
> I am interested in reconnect + fail over.

You have to carefully think about what you want to achieve and the obstacles
for doing a fail-over.

Things to think about:

Support fail-over of write operations? What happens if applications do
read-after-write or vice versa and there's a fail-over in the middle?

You have to maintain a list of functional servers. You have to implement a
strategy when to re-add a formerly failing server to the list of functional
servers.

etc.

Ciao, Michael.