[Gluster-devel] Re: Timeout settings and self-healing ? (WAS: HA failover test unsuccessful (inaccessible mountpoint))

[Krishna Srinivas]

On Mon, Apr 21, 2008 at 5:28 PM, Daniel Maher <dma+gluster at witbe.net> wrote:
> On Mon, 21 Apr 2008 17:20:53 +0530 "Krishna Srinivas"
>  <krishna at zresearch.com> wrote:
>
>  > Guido,
>  >
>  > Can you give the setup details, conf files?
>  > you can use http://glusterfs.pastebin.com for pasting conf files.
>
>  The following message to the mailing list on 02 April 2008 contains the
>  information you seek :
>
>  http://lists.nongnu.org/archive/html/gluster-devel/2008-04/msg00010.html
>

One thing common with all the three setups is that unify is being used on the
server side.

* using unify on the server side is not advisable, it should be used on the
  client side for most of the practical purposes. It should still work
  when used on the server side, but it is not working as expected (from your
  setup) This will be fixed.

* I have removed link from our wiki docs to
http://www.gluster.org/docs/index.php/GlusterFS_1.3_High_Availability_Storage_with_GlusterFS
  As this has unify on the server side without serving any purpose.
  (It was a user contributed example)

* The following setup is a good example:
http://www.gluster.org/docs/index.php/Simple_High_Availability_Storage_with_GlusterFS_1.3

* Loading AFR on the server side is as OK as loading it on client side. But as
  mentioned before, unify should be loaded on the client side for all
practical setups.
  Hence you guys should change your setup and use unify on client.

Thanks
Krishna

>
>  --
>  Daniel Maher <dma AT witbe.net>
>
>
>  _______________________________________________
>
>
> Gluster-devel mailing list
>  Gluster-devel at nongnu.org
>  http://lists.nongnu.org/mailman/listinfo/gluster-devel
>