Re: Replication delay

[Ángel L. Mateo <amateo@um.es>]

El 25/3/19 a las 19:56, Quanah Gibson-Mount escribió:
> None of your configurations set a keepalive.  If you have any sort of 
> load balancer/switch in between, they will often cut idle connections, 
> and replication will stop until slapd is restarted.
>
> I recommend always defining a keepalive interval in the syncrepl stanza 
> to avoid these types of issues.
>
> keepalive=240:10:30
>
> usually works well, unless the device is configured to close idle 
> connections at less than a 5 minute interval.
	Finally, it seems that the problem could be related with conections 
being closed, although I don't have any network device between servers.

	I have found messages like:

Apr  3 14:31:48 canis41 slapd[122414]: conn=211822 fd=21 closed 
(idletimeout)

	in my logs for replica connections.

	I tried to configure keepalive when you suggested it, but I configured 
it in just one connection of the replica.

	Now I have configured it in both consumers and since (yesterday) then I 
have no connections closed neither replica delays.

	Now I have to find why my system is closing these connections. The 
difference between the old farm and this new server is the ubuntu 
version (from 14.04 to 18.04). I don't know if systemd (used in ubuntu 
18.04) could be the reason.

--
Angel L. Mateo Martínez
Sección de Telemática
Área de Tecnologías de la Información
y las Comunicaciones Aplicadas (ATICA)
http://www.um.es/atica
Tfo: 868889150
Fax: 868888337