<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<br>
<br>
<div class="moz-cite-prefix">On 09/28/2015 04:15 PM, Pierre Léonard
wrote:<br>
</div>
<blockquote cite="mid:56091A33.4090705@agsic.net" type="cite">
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
<div class="moz-cite-prefix">Hi Soumya Koduri<br>
</div>
<blockquote cite="mid:5609194D.6040308@redhat.com" type="cite"> <br>
The only way I can imagine to avoid replication in case of
replicated volume is to kill one of the brick processes. <br>
CCin few who may be able to provide better solutions. <br>
<br>
Just asking out of curiosity, why do you want to disable
replication? To gain performance? <br>
</blockquote>
Yes and to get back storage. The usage planned two years ago have
changed, and we want suppres the replica which is to slow.<br>
<br>
</blockquote>
<br>
Ensure that there are no pending heals, split-brains etc. Then use
the remove-brick force command:<br>
`gluster volume remove-brick <volname> replica 1 host2:brick2
host4:brick4 .....host12:brick12 force`<br>
<br>
HTH,<br>
Ravi<br>
<blockquote cite="mid:56091A33.4090705@agsic.net" type="cite"> Many
thanks<br>
<br>
Pierre Léonard<br>
<br>
</blockquote>
<br>
</body>
</html>