<div dir="ltr"><div><div>Miklos, Thanks for this patch.<br><br></div>Team will review this, and update you on if this is good fix.<br><br></div>-Amar<br><div><div><br><br></div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, May 4, 2017 at 6:59 PM, Miklós Fokin <span dir="ltr"><<a href="mailto:miklos.fokin@appeartv.com" target="_blank">miklos.fokin@appeartv.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">
<p>Sorry, missing lines from the attachment.<br>
</p><div><div class="h5">
<br>
<div class="m_1538601115112990886moz-cite-prefix">On 05/04/2017 03:24 PM, Miklós Fokin
wrote:<br>
</div>
<blockquote type="cite">Hello,
<br>
<br>
I seem to have discovered what caused half of the problem.
<br>
I did update the bug report with a more detailed description, but
the short version is that the attached diff solves the issue when
we get an fstat with a size of 0 after killing a brick (not
letting the first update to fsync be from an arbiter).
<br>
My question is: should I make a review about it or should further
needed changes be investigated first?
<br>
<br>
Best regards,
<br>
Miklós
<br>
<br>
<br>
On 04/26/2017 12:58 PM, Miklós Fokin wrote:
<br>
<blockquote type="cite">Thanks for the response.
<br>
We didn't have the options set that the first two reviews were
about.
<br>
The third was about changes to performance.readdir-ahead.
<br>
I turned this feature off today with prefetch being turned on on
my computer, and the bug still appeared, so I would think that
the commit would not fix it either.
<br>
<br>
Best regards,
<br>
Miklós
<br>
<br>
<br>
On 04/25/2017 01:26 PM, Raghavendra Gowdappa wrote:
<br>
<blockquote type="cite">Recently we had worked on some patches
to ensure correct stats are returned.
<br>
<br>
<a class="m_1538601115112990886moz-txt-link-freetext" href="https://review.gluster.org/15759" target="_blank">https://review.gluster.org/<wbr>15759</a>
<br>
<a class="m_1538601115112990886moz-txt-link-freetext" href="https://review.gluster.org/15659" target="_blank">https://review.gluster.org/<wbr>15659</a>
<br>
<a class="m_1538601115112990886moz-txt-link-freetext" href="https://review.gluster.org/16419" target="_blank">https://review.gluster.org/<wbr>16419</a>
<br>
<br>
Referring to these patches and bugs associated with them might
give you some insight into the nature of the problem. The
major culprit was interaction between readdir-ahead and
stat-prefetch. So, the issue you are seeing might be addressed
by these patches.
<br>
<br>
----- Original Message -----
<br>
<blockquote type="cite">From: "Miklós Fokin"
<a class="m_1538601115112990886moz-txt-link-rfc2396E" href="mailto:miklos.fokin@appeartv.com" target="_blank"><miklos.fokin@appeartv.com></a>
<br>
To: <a class="m_1538601115112990886moz-txt-link-abbreviated" href="mailto:gluster-devel@gluster.org" target="_blank">gluster-devel@gluster.org</a>
<br>
Sent: Tuesday, April 25, 2017 3:42:52 PM
<br>
Subject: [Gluster-devel] fstat problems when killing with
stat prefetch turned on
<br>
<br>
Hello,
<br>
<br>
I tried reproducing the problem that Mateusz Slupny was
experiencing
<br>
before (stat returning bad st_size value on self-healing) on
my own
<br>
computer with only 3 bricks (one being an arbiter) on
3.10.0.
<br>
The result with such a small setup was that the bug appeared
both on
<br>
killing and during the self-healing process, but only rarely
(once in
<br>
hundreds of tries) and only with performance.stat-prefetch
turned on.
<br>
This might be a completely different issue as on the setup
Matt was
<br>
using, he could reproduce it with the mentioned option being
off, it
<br>
always happened but only during recovery, not after killing.
<br>
I did submit a bug report about this:
<br>
<a class="m_1538601115112990886moz-txt-link-freetext" href="https://bugzilla.redhat.com/show_bug.cgi?id=1444892" target="_blank">https://bugzilla.redhat.com/<wbr>show_bug.cgi?id=1444892</a>.
<br>
<br>
The problem is as Matt wrote is that this causes data
corruption if one
<br>
is to use the returned size on writing.
<br>
Could I get some pointers as to what parts of the gluster
code I should
<br>
be looking at to figure out what the problem might be?
<br>
<br>
Thanks in advance,
<br>
Miklós
<br>
<br>
______________________________<wbr>_________________
<br>
Gluster-devel mailing list
<br>
<a class="m_1538601115112990886moz-txt-link-abbreviated" href="mailto:Gluster-devel@gluster.org" target="_blank">Gluster-devel@gluster.org</a>
<br>
<a class="m_1538601115112990886moz-txt-link-freetext" href="http://lists.gluster.org/mailman/listinfo/gluster-devel" target="_blank">http://lists.gluster.org/<wbr>mailman/listinfo/gluster-devel</a>
<br>
</blockquote>
</blockquote>
<br>
</blockquote>
<br>
<br>
<fieldset class="m_1538601115112990886mimeAttachmentHeader"></fieldset>
<br>
<pre>______________________________<wbr>_________________
Gluster-devel mailing list
<a class="m_1538601115112990886moz-txt-link-abbreviated" href="mailto:Gluster-devel@gluster.org" target="_blank">Gluster-devel@gluster.org</a>
<a class="m_1538601115112990886moz-txt-link-freetext" href="http://lists.gluster.org/mailman/listinfo/gluster-devel" target="_blank">http://lists.gluster.org/<wbr>mailman/listinfo/gluster-devel</a></pre>
</blockquote>
<br>
</div></div></div>
<br>______________________________<wbr>_________________<br>
Gluster-devel mailing list<br>
<a href="mailto:Gluster-devel@gluster.org">Gluster-devel@gluster.org</a><br>
<a href="http://lists.gluster.org/mailman/listinfo/gluster-devel" rel="noreferrer" target="_blank">http://lists.gluster.org/<wbr>mailman/listinfo/gluster-devel</a><br></blockquote></div><br><br clear="all"><br>-- <br><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div>Amar Tumballi (amarts)<br></div></div></div></div></div>
</div>