Quanah Gibson-Mount quanah@symas.com schrieb am 25.03.2019 um 19:56 in
Nachricht <CF0FD0042D3D17C49C943E62@[192.168.1.39]>:
--On Monday, March 25, 2019 9:34 AM +0100 "Ángel L. Mateo" amateo@um.es wrote:
- I have a farm with syncrepl running for years without any problem. This
farm is running 2.4.43. Servers of this farm are named canis3x
- I have a new farm with servers named canis41 and canis42 (with 2.4.47),
where: ** canis41 is syncrepl with canis39 (to keep both farms synchronized) ** canis41 and canis42 has delta-syncrepl between them (my idea is to switch off canis3x farm and then left just this delta-syncrepl)
The problem I'm having is with syncrepl between canis39 and canis41.
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.
Hi!
I don't understand that: Keepalive (as I know it) has nothing to do with idle connections, but only with dead connections. And what's the magic with 5 minutes?
Regards, Ulrich
--Quanah
--
Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: http://www.symas.com