<html><head>
<meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type">
</head><body bgcolor="#FFFFFF" text="#000000">Using the virt group which
does-<br>
<br>
cluster.quorum-type: auto<br>
cluster.server-quorum-type: server<br>
<br>
And from my understanding this means that quorum enforcement is enabled,
and that it's brick related so if I only have two bricks then I cannot
achieve 51% and split brain is likely even with a third node in the
pool. If I move to fixed 1, then the likelihood of splitbrain is high.<br>
<br>
So what do I do here, get a fourth node and do distributed replicated?
Trying to find the best performance option while minimizing split brain
potential and I'm stuck with 3.6 clients ( RHEV-H ) so the
aforementioned arbiter is out, Replica 3 just seems to be a pig on the
client side due to the triple write.<br>
<br>
<div class="moz-signature">-- <br>Scott H.<br><br>
<br>
</div>
<br>
<span>
</span><br>
<blockquote style="border: 0px none;"
cite="mid:55A693BA.9020004@redhat.com" type="cite">
<div style="margin:30px 25px 10px 25px;" class="__pbConvHr"><div
style="display:table;width:100%;border-top:1px solid
#EDEEF0;padding-top:5px"> <div
style="display:table-cell;white-space:nowrap;vertical-align:middle;width:100%">
<a moz-do-not-send="true" href="mailto:ravishankar@redhat.com"
style="color:#737F92
!important;padding-right:6px;font-weight:bold;text-decoration:none
!important;">Ravishankar N</a></div> <div
style="display:table-cell;white-space:nowrap;vertical-align:middle;">
<font color="#9FA2A5"><span style="padding-left:6px">July 15, 2015 at
1:09 PM</span></font></div> </div></div>
<div style="color:#888888;margin-left:24px;margin-right:24px;"
__pbrmquotes="true" class="__pbConvBody">
<br>
<br>On 07/15/2015 09:22 PM, Scott Harvanek wrote:
<br><blockquote type="cite">I saw Ravi mention something earlier about
the new arbiter volumes in 3.7
<br>
<br>- Can you run a arbiter 3.7 server volume with 3.6 clients? Or does
this not work?
<br>
<br>
<br></blockquote>
<br>Unfortunately, no. This is because the logic for the feature is
mostly
in the AFR translator which is loaded on the client process (with some
minor logic in the arbiter translator loaded on the brick process). So
both clients and servers need to be 3.7.
<br>
<br>
<br><blockquote type="cite">Quorum question -
<br>
<br>- 3.6 server and clients, replica 2, however there's three+ servers
in
the cluster ( gluster pool list ). If one of the replicas fail is
quorum maintained via the other servers in the cluster even though
they are not part of the volume?
<br>
<br></blockquote>That would depend on what type of quorum is enabled
(client or server
side). Each type has a different impact. You can look for the meaning of
cluster.quorum-type, cluster.quorum-count (these are client side quorum
options, applicable only to replicate volumes) and
cluster.server-quorum-type, cluster.server-quorum-ratio (these are
server side quorum options; they are not specific to replicate volumes)
in 'gluster volume set help`.
<br>
<br>HTH,
<br>Ravi
<br></div>
<div style="margin:30px 25px 10px 25px;" class="__pbConvHr"><div
style="display:table;width:100%;border-top:1px solid
#EDEEF0;padding-top:5px"> <div
style="display:table-cell;white-space:nowrap;vertical-align:middle;width:100%">
<a moz-do-not-send="true" href="mailto:scott.harvanek@login.com"
style="color:#737F92
!important;padding-right:6px;font-weight:bold;text-decoration:none
!important;">Scott Harvanek</a></div> <div
style="display:table-cell;white-space:nowrap;vertical-align:middle;">
<font color="#9FA2A5"><span style="padding-left:6px">July 15, 2015 at
11:52 AM</span></font></div> </div></div>
<div style="color:#888888;margin-left:24px;margin-right:24px;"
__pbrmquotes="true" class="__pbConvBody">I saw Ravi mention something
earlier about the new arbiter volumes in 3.7
<br>
<br>- Can you run a arbiter 3.7 server volume with 3.6 clients? Or does
this not work?
<br>
<br>
<br>Quorum question -
<br>
<br>- 3.6 server and clients, replica 2, however there's three+ servers
in
the cluster ( gluster pool list ). If one of the replicas fail is
quorum maintained via the other servers in the cluster even though they
are not part of the volume?
<br>
<br></div>
</blockquote>
<br>
</body></html>