[Systems] Clustering Remora
dfarning at sugarlabs.org
Fri Aug 28 21:32:27 EDT 2009
On Fri, Aug 28, 2009 at 6:29 PM, Bernie Innocenti<bernie at codewiz.org> wrote:
> El Fri, 28-08-2009 a las 06:15 -0500, David Farning escribió:
>> a.sl.o handles the syncing. Before the bundle is marked public, via
>> the editor panel, it is stored locally. After being marked public, it
>> is automatically pushed to the download server. All syncing and
>> redirecting is handled by Remora.
>> The big win is that by moving the files off the primary server it
>> reduces cache churn on the primary servers and db servers.
Are we going to use cacheboy as our primary download mechanism? If
so, it looks like we can point activities.sl.o to use
sugarlabs.cdn.cacheboy.net. (the origin cacheboy server) From there,
the activities would be served via the cacheboy system.
Does this mean running two separate cacheboy nodes, an origin node and
a mirror node? How is this generally handled?
>> +1, Whenever you get a chance wes and I can start work on this.
>> possible work flow:
>> 1. Bernie (or vm maintainer) creates new vms.
>> a. activites-devel.sl.o
>> b. activites-devel-dl.sl.o
>> 2. Bernie (or dns maintainer) points dns names at new vms.
>> 3. Wes (or aslo admin) sets up stock also server - I can help point
>> wes to the necessary http configs and create db backups.
>> 4. Wes (or aslo admin) syncs aslo instances against correct git branch
>> - will appreciate alsroots help here.
>> repeat 1-4 with activites-testing.sl.o and activites-testing-dl sl.o
>> repeat 1-4 with activites-dl.sl.o -- and go live
> Ok. It will take me a while, but I'll definitely work on it high
> // Bernie Innocenti - http://codewiz.org/
> \X/ Sugar Labs - http://sugarlabs.org/
More information about the Systems