Reposting this request from Davide Bozzelli as a new post/thread, since it is most deserving of an answer.
Although I would add... A simple 'tutorial' for how to monitor your system and how to interpret and/or translate what you are monitoring into potential values for the variables in question. Ie, is it only RAM usage? If so, what exactly is 'normal'? And what is 'abnormal', and which setting should be tweaked and when? In my opinion, whatever the default values for these are should work out of the box for a reasonably sized organization (20-100 users?)... ******************************************************* Given the high demand about this topic, could u please publish some "best practices" about tuning this values ? Something like "if you have 50 users then set them with these values" or something like. It's not intuitive as you might think . Thx Il 22/01/15 17:47, Ludovic Marcotte ha scritto: > On 22/01/2015 10:38, lis...@outlook.de wrote: >> Any ideas concerning the reason? > Have you tuned your proxy read timeouts in nginx? You had 90 seconds > from your previous post, which is way too low. > > Have you also tuned: > > SOGoMaximumPingInterval > SOGoMaximumSyncInterval > SOGoInternalSyncInterval > > If not, a sogod process will not keep the EAS connection active for > more than 10 to 30 seconds - which will force your EAS device to > re-establish the connection, over and over. > > If you increase these values, you must also increase your nginx/Apache > timeouts because they won't receive a response from sogod for X amount > of minutes. > > If you do that, you must also adjust WOWatchDogRequestTimeout, because > by default, sogod processes aren't allowed to take more than 10 > minutes to handle a request. > -- users@sogo.nu https://inverse.ca/sogo/lists -- users@sogo.nu https://inverse.ca/sogo/lists
