Hello We have several instances. One with postgres locally and EAS, as in your setup. It has 48 GB of RAM and 200 Worker. It didn't have any problems with OOM Killer since half a year.
But separating EAS clients from the rest of the users is a good idea. When using multiple SOGo instances, you need one shared memcached server.In our tests we only did use IP pinning for distributing users between instances.
Some users switched those within sessions by roaming around.Because of that we had to also share NGMimeBuildMimeTempDirectory and SOGoMailSpoolPath via NFS.
Kind regards, Christian MackAm 04.06.25 um 15:52 schrieb "Alexander Mischke" ([email protected]):
Hello, does anybody here use SOGo in a clustered setup? If so, what is your setup and which recommendations do you have? We currently run SOGo with 40 workers and EAS on a single VM (24 GB RAM, 16 Cores), postgres runs locally on the server. The OOM Killer is getting active every now and then, so we would like to spread out the workloads to 2 servers (one for regular SOGo web and one for EAS). Thanks and best regards, Alexander Mischke --
-- Christian Mack Universität Konstanz Kommunikations-, Informations-, Medienzentrum (KIM) Abteilung IT-Dienste Forschung, Lehre, Infrastruktur 78457 Konstanz +49 7531 88-4416
smime.p7s
Description: Kryptografische S/MIME-Signatur
