From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-bw0-f46.google.com ([209.85.214.46]:47019 "EHLO mail-bw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751003Ab1FOFe7 (ORCPT ); Wed, 15 Jun 2011 01:34:59 -0400 Received: by bwz15 with SMTP id 15so180523bwz.19 for ; Tue, 14 Jun 2011 22:34:58 -0700 (PDT) Message-ID: <4DF8447C.40409@gmail.com> Date: Wed, 15 Jun 2011 11:04:52 +0530 From: Amit Tewari | gmail To: linux-nfs@vger.kernel.org Subject: FSCache and NFS Attribute caching Content-Type: text/plain; charset=ISO-8859-1; format=flowed Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 Hi, fscache is being considered to be deployed with NFS 3 server in a metadata rich environment (millions of small files, with very low rate of file data change). At present actimeo=600 is used with NFS mount options (without fscache), to cache file attributes. I am not clear on 1. How FSCache handles meta data caching? 2. Does FSCache consider actimeo values before making a request to server? 3. When FSCache has to do metadata lookup from server, does it check local cache first and then server lookup is performed? need some definitive answer. Nearest that i could find anything is: http://kerneltrap.org/Linux/Local_Caching_For_Network_Filesystems http://linux-nfs.org/pipermail/nfsv4/2006-August/004844.html (for NFS v4) http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/fscachelimitnfs.html Ultimate goal is to reduce metadata lookup load from the NFS server. Any clarity/ pointers/ help will be highly appreciated. Thanks in advance. Regards Amit