[Gluster-devel] Glusterfs HA & AFR

[Chris Mavin]

Hi All,
 
Can someone clarify the following for me.
 
I'm using Glusterfs RC 2.
 
If I am using the HA translator rather than AFR, Does this expect to be using a shared storage accessible from both servers?
 
Put another way, is the HA translator simply a way of providing High availability for a SAN?
 
When I test it using 2 servers the active server is written to fine but the backup server simply gets a 0 byte file written to it.
 
Thanks
 
Chris.


> Date: Mon, 16 Feb 2009 16:41:50 +0000
> From: address@hidden
> To: address@hidden
> Subject: Re: [Gluster-devel] File Clobbering Bug
>
> Sorry, forgot to mention - no performance translators are being used.
>
> On Mon, 16 Feb 2009 11:25:44 +0530, Krishna Srinivas
> <address@hidden>
> wrote:
> > Gordan,
> >
> > More clues would help us.
> >
> > * Can you reproduce the problem without samba?
> > * Are perf xlators loaded? can you mail the vol files?
> > * do the logs indicate anything?
> >
> > Krishna
> >
> > On Mon, Feb 16, 2009 at 6:05 AM, Gordan Bobic <address@hidden> wrote:
> >> I seem to have stumbled upon another bug.
> >>
> >> Two Linux servers in AFR (2.0.0rc1), one client that exports the
> >> directory
> >> via samba.
> >>
> >> If one server is shut down, and a client connects to the client and
> >> modifies
> >> a file via samba, the changes get clobbered when the 2nd server comes
> >> back.
> >> When the server that was disconnected rejoins it seems to take
> precedence
> >> and clobbers the changes made while it was down.
> >>
> >> Gordan
> >>
> >>
> >> _______________________________________________
> >> Gluster-devel mailing list
> >> address@hidden
> >> http://lists.nongnu.org/mailman/listinfo/gluster-devel
> >>
>
>
> _______________________________________________
> Gluster-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/gluster-devel