<div dir="ltr"><div>So, just to recap, is it ok to clone repo from github, go to tag 3.3.2qa3, stop glusterd, configure, make & make install?<br><br></div>Regards,<br>Stefano<br><div><br></div></div><div class="gmail_extra">
<br><br><div class="gmail_quote">On Mon, Jun 3, 2013 at 5:54 PM, Vijay Bellur <span dir="ltr"><<a href="mailto:vbellur@redhat.com" target="_blank">vbellur@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 06/02/2013 01:42 PM, Stefano Sinigardi wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Also directories got removed. I did a really bad job in that script,<br>
wrong sed and path was not truncated and replaced with the fuse<br>
mountpoint...<br>
Yes I can move to 3.3.2qa3. At the moment I have gluster installed as<br>
per the semiosis repository. What is the best way for me to move to this<br>
other version (I think that I have to build it from source?)?<br>
</blockquote>
<br></div>
I think qa builds are not part of semiosis' ppa repository. Copying him to chime in if it is otherwise.<br>
<br>
If the builds are not available, you will need to build it from source.<br>
<br>
Regards,<br>
Vijay<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">
Thanks and best regards,<br>
Stefano<br>
<br>
<br>
On Sun, Jun 2, 2013 at 4:15 PM, Vijay Bellur <<a href="mailto:vbellur@redhat.com" target="_blank">vbellur@redhat.com</a><br></div><div class="im">
<mailto:<a href="mailto:vbellur@redhat.com" target="_blank">vbellur@redhat.com</a>>> wrote:<br>
<br>
On 06/02/2013 11:35 AM, Stefano Sinigardi wrote:<br>
<br>
Dear Vijay,<br>
the filesystem is ext4, on a GPT structured disk, formatted by<br>
Ubuntu 12.10.<br>
<br>
<br>
A combination of ext4 on certain kernels and glusterfs has had its<br>
share of problems<br></div>
(<a href="https://bugzilla.redhat.com/__show_bug.cgi?id=838784" target="_blank">https://bugzilla.redhat.com/_<u></u>_show_bug.cgi?id=838784</a><br>
<<a href="https://bugzilla.redhat.com/show_bug.cgi?id=838784" target="_blank">https://bugzilla.redhat.com/<u></u>show_bug.cgi?id=838784</a>>) for readdir<div><div class="h5"><br>
workloads. I am not sure if the Ubuntu 12.10 kernel is affected by<br>
this bug as well. GlusterFS 3.3.2 has an improvement which will<br>
address this problem seen with ext4.<br>
<br>
<br>
The rebalance I did was with the command<br>
gluster volume rebalance data start<br>
but in the log it got stuck on a file that I cannot remember (was a<br>
small working .cpp file, saying that it was going to be moved to<br>
an much<br>
more occupied replica, and it repeated this message until<br>
writing a log<br>
that was a few GB).<br>
Then I stopped it and restarted with<br>
gluster volume rebalance data start force<br>
in order to get rid of this problems about files going to bricks<br>
already<br>
highly occupied.<br>
Because I was almost stuck, remembering that a rebalance solved<br>
another<br>
problem I had as a miracle, I retried it, but got stuck in a<br>
.dropbox-cache folder. That is not a very important folder, so I<br>
thought<br>
I could remove it. I launched a script to find all the files<br>
looking at<br>
all the bricks but removing them from the fuse mountpoint. I<br>
don't know<br>
what went wrong (the script is very simple, the problem maybe<br>
was that<br>
it was 4 am in the night) but the fact is that files got removed<br>
calling<br>
rm at the bricks mountpoints, not the fuse one. So I think that<br>
now I'm<br>
in a even worse situation that before. I just stopped working on it,<br>
asking for some time from my colleagues (at least data is still<br>
there,<br>
on the bricks, just sparse on all of them) in order to think<br>
well about<br>
how to proceed (maybe destroying it and rebuilding it, but it<br>
will be<br>
very time consuming as I don't have so much free space elsewere<br>
to save<br>
everything, also it's very difficult to save from the fuse<br>
mountpoint as<br>
it's not listing all the files)<br>
<br>
<br>
Were only files removed from the brick mountpoints or did<br>
directories get removed too? Would it be possible for you to move<br>
to 3.3.2qa3 and check if ls does list all files present in the<br>
bricks? Note that, qa3 is not yet GA and might see a few fixes<br>
before it becomes so.<br>
<br>
Regards,<br>
Vijay<br>
<br>
<br>
<br>
</div></div></blockquote>
<br>
</blockquote></div><br></div>