<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>Hello</p>
<p>I am using Glusterfs 3.10.5 on CentOS7. A replicated distributed
volume with a dist-rep hot tier.</p>
<p>During data migration, we noticed the tierd.log on one of nodes
was huge. Upon review it seemed to be stuck on a certain set of
files. Running a "gluster vol heal VOL info" showed that those
same files caused problems in the tier, were in split brain.</p>
<p> So we went to fix split brain, we used "source-brick" and chose
each of the bricks as the master to heal from. <br>
</p>
<blockquote>
<p>gluster vol heal VOL split-brain source-brick
HOSTNAME:/brick/dir<br>
</p>
</blockquote>
<p>However when we ran the heal command above, It spit out that each
file was not in split brain and that it healed 0.</p>
<p>We then double checked the status of the files listed in heal
info output with :</p>
<blockquote>
<p>getfattr -n replica.split-brain-status <path-to-file></p>
</blockquote>
<p>And it reported that the files were not in split brain....</p>
<p>Any insight would be greatly appreciated.</p>
<p>Brett<br>
</p>
<blockquote>
<p> <br>
</p>
</blockquote>
</body>
</html>