Is there any pointers on how to successfully apply the gluster replicated setup for sharing files without suffering from bad performance for normal LAMP applications... On all servers I would have a brick with data, in replicated setup. LAMP normally does frequent reads and infrequent writes.. I am after a way to maximize the read performance on a gluster mount while the writes can be even slow... it does not matter very much since it is infrequent.. I have verified a working split between reads and writes thanx to great help from Franciscis in thread "Separating reads and writes for performance benefit" but some applications just cannot cope with split reads and writes.. So, I am looking for a way to obtain a "normal-filesystem-like" read performance while write can be much less efficient. It neither very importain that the read copy is always verified the latest version existing. thanx in advance for any pointers -- -- Haris Zukanovic
