More to the point the first link doesn't 100% cover HA. The roadmap for
1.4 adds in the cient side the HA module that will complete this
configuration.
In that wiki link we built an HA Server Cluster where all replication,
self-heal etc happens on the servers not the clients. With this the
client still needs mount point redundancy. Once the HA translator is
complete this config will actually become useful.
Daniel van Ham Colchete wrote:
Hi Jordi,
I'm sorry but please don't read the first link at first. It's a more
advanced configuration. The second link will suffice.
Best regards,
Daniel Colchete
On 7/10/07, Daniel van Ham Colchete <daniel.colchete@xxxxxxxxx> wrote:
Hi Jordi,
I'm not a Gluster developer and I think you might find the following
links
very helpfull:
http://www.gluster.org/docs/index.php/GlusterFS_High_Availability_Storage_with_GlusterFS
http://www.gluster.org/docs/index.php/GlusterFS_Translators#Automatic_File_Replication_Translator
Best regards,
Daniel Colchete
On 7/10/07, Jordi Prats <jprats@xxxxxxxx> wrote:
>
> Hi,
> I found that with a single node failure the entire filesystem becomes
> unusable with this error:
>
> [root@cat02 tmp]# ls
> ls: .: Transport endpoint is not connected
>
> There's any way to avoid this error so it just goes on loosing the
files
>
> of the failed node? If not, there's any plan to implement it on future
> releases?
>
> Finally, there's any way to give some redundancy to glusterFS so a
> single file is on two nodes? This could be also a performance
> improvement since the same data can be read from two nodes at once.
>
> Thank you!
> Jordi
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel@xxxxxxxxxx
> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>
_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
http://lists.nongnu.org/mailman/listinfo/gluster-devel
