Re: Cluster takes around 48 hours to synchronize after deletion

[Quanah Gibson-Mount <quanah@symas.com>]

--On Wednesday, June 26, 2019 5:16 PM +0000 x0101 <x0101@protonmail.com> 
wrote:

> We have an OpenLDAP 2.4 cluster of three nodes configured in multi-master
> and accessed through a VIP in round-robin. The three machines run RHEL7.
>
> What could be the cause and what could we do to further troubleshoot the
> issue?  Thanks in advance.

a) Don't use the outdated builds from RedHat.  You generally have the 
following choices here:
  1) Build OpenLDAP yourself
  2) Use a free replacement, such as Symas' OpenLDAP for Linux 
(<https://repo.symas.com/sofl/rhel7/>) or the LTB project's OpenLDAP build 
(<https://ltb-project.org/documentation/openldap-rpm#yum_repository>)
  3) Obtain a commercially supported version of OpenLDAP, such as Symas 
OpenLDAP Gold (<https://symas.com/symasopenldap/>.  There is also optional 
commerical support for Symas OpenLDAP for Linux

b) Use delta-syncrepl, not standard syncrepl.

c) Ensure that in addition to stats logging you also have sync logging 
enabled, otherwise you'll never be able to tell what's occuring.

Also, you fail to note your configuration settings for the syncprov 
overlay, which is rather critical to note as well.  For example, you may 
have a low sessionlog value set, which can cause all sorts of havoc with 
standard syncrepl, especially in older releases.

Regards,
Quanah

--

Quanah Gibson-Mount
Product Architect
Symas Corporation
Packaged, certified, and supported LDAP solutions powered by OpenLDAP:
<http://www.symas.com>