<!DOCTYPE html>
<html>
<head>
<title></title>
</head>
<body><div>On Tue, Jun 16, 2015, at 10:29 PM, Tecnologia Charne.Net wrote:<br></div>
<blockquote type="cite"><div>
Problem was btrfs metadata:<br></div>
<div> </div>
<div>
cyrus squatter <a href="mailto:user/javier@unnoba.edu.ar">user/javier@unnoba.edu.ar</a><br></div>
<div>
Writing index update: No space left on device<br></div>
<div> </div>
<div> </div>
<div>
Filesystem checks:<br></div>
<div> </div>
<div>
mda:~# btrfs filesystem df /var/spool/cyrus/<br></div>
<div>
Data, single: total=1.26TiB, used=349.31GiB<br></div>
<div>
System, DUP: total=8.00MiB, used=160.00KiB<br></div>
<div>
System, single: total=4.00MiB, used=0.00B<br></div>
<div> <b>Metadata, DUP: total=3.50GiB, used=2.99GiB</b><br></div>
<div>
Metadata, single: total=8.00MiB, used=0.00B<br></div>
<div> </div>
<div> </div>
<div>
btrfs filesystem usage /var/spool/cyrus<br></div>
<div>
Overall:<br></div>
<div>
Device size: 1.27TiB<br></div>
<div>
Device allocated: 1.27TiB<br></div>
<div>
Device unallocated: 60.00KiB<br></div>
<div>
Device missing: 0.00B<br></div>
<div>
Used: 355.30GiB<br></div>
<div>
Free (estimated): 943.69GiB (min: 943.69GiB)<br></div>
<div>
Data ratio: 1.00<br></div>
<div>
Metadata ratio: 2.00<br></div>
<div>
Global reserve: 0.00B (used: 0.00B)<br></div>
<div> </div>
<div>
Data,single: Size:1.26TiB, Used:349.31GiB<br></div>
<div>
/dev/xvdb1 1.26TiB<br></div>
<div> </div>
<div>
Metadata,single: Size:8.00MiB, Used:0.00B<br></div>
<div>
/dev/xvdb1 8.00MiB<br></div>
<div> </div>
<div> <b>Metadata,DUP: Size:3.50GiB, Used:2.99GiB</b><b><br></b><b> /dev/xvdb1 7.00GiB</b></div>
<div> </div>
<div>
System,single: Size:4.00MiB, Used:0.00B<br></div>
<div>
/dev/xvdb1 4.00MiB<br></div>
<div> </div>
<div>
System,DUP: Size:8.00MiB, Used:160.00KiB<br></div>
<div>
/dev/xvdb1 16.00MiB<br></div>
<div> </div>
<div>
Unallocated:<br></div>
<div>
/dev/xvdb1 60.00KiB<br></div>
<div> </div>
<div> </div>
<div>
# btrfs balance start -v -dusage=5 /var/spool/cyrus/<br></div>
<div>
Dumping filters: flags 0x1, state 0x0, force is off<br></div>
<div>
DATA (flags 0x2): balancing, usage=5<br></div>
<div>
Done, had to relocate 213 out of 1302 chunks<br></div>
<div> </div>
<div> </div>
<div>
Now, it's working.<br></div>
</blockquote><div> </div>
<div>Why would anybody ever run production data on btrfs?<br></div>
<div> </div>
<div><a href="https://btrfs.wiki.kernel.org/index.php/Main_Page">https://btrfs.wiki.kernel.org/index.php/Main_Page</a><br></div>
<div> </div>
<div>"The Btrfs code base is under heavy development."<br></div>
<div> </div>
<div><a href="https://btrfs.wiki.kernel.org/index.php/Gotchas">https://btrfs.wiki.kernel.org/index.php/Gotchas</a><br></div>
<div> </div>
<div>"Files with a lot of random writes can become heavily fragmented
(10000+ extents) causing trashing on HDDs and excessive multi-second
spikes of CPU load on systems with an SSD or large amount a RAM."<br></div>
<div> </div>
<div>Unfortunately, both cyrus.index files and skiplist/twoskip databases get a lot of random writes - though both repack occasionally which reduces fragmentation.<br></div>
<div> </div>
<div>I've tried other filesystems a few times, and I keep coming back to ext4 for consistent performance and crash safety (at least on decent hardware which doesn't lie about fsync) - XFS has lost data, btrfs has always been shiny at first, but turned into a performance disaster over time.<br></div>
<div> </div>
<div id="sig567075"><div class="signature">--<br></div>
<div class="signature">Bron Gondwana<br></div>
<div class="signature">brong@fastmail.fm<br></div>
<div class="signature"> </div>
</div>
<div> </div>
</body>
</html>