mahara-contributors team mailing list archive

[PiersHarding <piers@xxxxxxxxx>]

Hi John - 
It does sound a bit like an AD issue.  If you look at it from a standard high availability LDAP point of view - usually if you have a fail-over, the IP address for the LDAP server is a floating one, that is then rotated to the recovery host.  So - from a service perspective - the service has never gone away, or was only temporarily unavailable (while the fail-over switch was performed).  I think it is reasonable to expect that a back up directory is a mirror, and there fore has identical data in terms of directory structure and accounts etc.  Is it not possible to use active directory replication between the two servers, and then maybe use something like  the firewall to load balance across them (I know you can do this with CheckPoint)?  That might be a way round this?

Cheers,
Piers Harding.

-- 
You received this bug notification because you are a member of Mahara
Contributors, which is subscribed to Mahara.
https://bugs.launchpad.net/bugs/855525

Title:
  Logon failure - LDAP authentication tied to one server

Status in Mahara ePortfolio:
  Won't Fix

Bug description:
  Mahara 1.4.0
  Linux Centos 5.7
  MySQL
  All browsers

  User logons failing when username and password are correct.

  We added two new Microsoft AD servers to our institution. However, all
  accounts in this institution cannot logon using these servers due to
  the users having the auth_instance declared in their usr entry
  (authinstance). If I update the auth_instance with the new server
  details they can logon. This means if that server fails users will not
  be able to logon even though we have other servers listed which can
  authenticate.

To manage notifications about this bug go to:
https://bugs.launchpad.net/mahara/+bug/855525/+subscriptions