<div dir="ltr"><div><div><div>Tested-by: Valerio Pachera<br><br></div>Result of my testing is that snapshots get correctly removed freeing space in the cluster.<br></div><div>I've been testing full replication (-c 2) and erasure code (-c 2:1).<br>
</div><br>Removing all snapshots and vdi (so no vdi left in the cluster), most of the time leave almost no objects (e.g. 6M).<br></div><div>In some case (but we can't actually reproduce it), more objects are left (e.g. 600M).<br>
<br></div><div>Here's an example<br><br>s test 1 5.0 GB 1.0 GB 0.0 MB 2014-05-21 12:46 7c2b25 2 <br>s test 2 5.0 GB 4.4 GB 340 MB 2014-05-21 12:46 7c2b26 2 <br>
s test 3 5.0 GB 24 MB 4.7 GB 2014-05-21 12:49 7c2b27 2 <br>s test 4 5.0 GB 4.4 GB 340 MB 2014-05-21 12:49 7c2b28 2 <br>s test 5 5.0 GB 28 MB 4.7 GB 2014-05-21 12:52 7c2b29 2 <br>
s test 6 5.0 GB 32 MB 4.7 GB 2014-05-21 12:52 7c2b2a 2 <br> test 0 5.0 GB 0.0 MB 4.7 GB 2014-05-21 12:53 7c2b2b 2 <br><br>This is total size (dog node info), before removing the snapshot; after removing snapshots one after the other and, finally, the vdi.<br>
Total 682 GB 20 GB 663 GB 2%<br>Total 682 GB 18 GB 664 GB 2%<br>Total 682 GB 18 GB 664 GB 2%<br>Total 682 GB 9.5 GB 673 GB 1%<br>Total 682 GB 9.5 GB 673 GB 1%<br>Total 682 GB 9.4 GB 673 GB 1%<br>
Total 682 GB 9.4 GB 673 GB 1%<br>Total 682 GB 0.1 MB 682 GB 0%<br><br></div><div>Sheepdog daemon version 0.8.0_171_g4414730<br></div><br></div>