|
From: | Xavier Hernandez |
Subject: | Re: [Gluster-devel] [RFC] A new caching/synchronization mechanism to speed up gluster |
Date: | Thu, 06 Feb 2014 09:51:43 +0100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0 |
Hi Avati, El 06/02/14 00:24, Anand Avati ha escrit: Yes, that is a basic requirement for many features. I saw the client_t changes but haven't had time to see if they could be used to implement the kind of mechanism I proposed. This will need a look. When I started implementing the DFC translator (https://forge.gluster.org/disperse/dfc) I needed something very similar but at that time there wasn't any suitable client_t implementation I could use. I solved it by using a pool of special getxattr requests that the translator on the bricks stores until it needs to send some message back to the client. It's not a great solution but it works with the available resources at the moment. I have some ideas on how to implement it and some special cases, but I need to work more on it before it can be considered a valid model. I just wanted to propose the idea to see if it could be valid or not before spending too much of my scarce time working on it. I'll try to get a more detailed picture to discuss it. Best regards, Xavi
|
[Prev in Thread] | Current Thread | [Next in Thread] |