Hi All,<div><br></div><div>Testing sheepdog.</div><div><br></div><div>create cluster for 3node with --copies=2<div>created vdi with 2 gb. start VM write date to vdi <br clear="all">dd if=/dev/zero of=/dev/vda bs=1M count=2000</div>
<div><br></div><div><div>collie node info</div><div>Id Size Used Use%</div><div> 0 20 GB 740 MB 3%</div><div> 1 20 GB 636 MB 3%</div><div> 2 4.8 GB 632 MB 12%</div><div>Total 44 GB 2.0 GB 4%, total virtual VDI Size 2.0 GB</div>
<div><br></div><div>so far so good.</div><div><br></div><div>now simulate node failure ( on node3 by killling sheep )</div><div>after recovery complited:</div><div><div>collie node info</div><div>Id Size Used Use%</div>
<div> 0 20 GB 2.0 GB 10%</div><div> 1 19 GB 1.9 GB 9%</div><div>Total 39 GB 3.8 GB 9%, total virtual VDI Size 2.0 GB</div></div><div><br></div><div>start sheep on 3 node again - wait for recovery, after finished:</div>
<div><div>collie node info</div><div>Id Size Used Use%</div><div> 0 19 GB 1.4 GB 7%</div><div> 1 19 GB 1.2 GB 6%</div><div> 2 3.5 GB 1.2 GB 34%</div></div><div>Total 41 GB 3.9 GB 9%, total virtual VDI Size 2.0 GB</div>
<div><br></div><div>one more failure on the same node and bring it back again:</div><div><div>collie node info</div><div>Id Size Used Use%</div><div> 0 18 GB 1.4 GB 7%</div><div> 1 18 GB 1.2 GB 6%</div>
<div> 2 2.3 GB 1.2 GB 53%</div><div><br></div><div>Total 39 GB 3.9 GB 10%, total virtual VDI Size 2.0 GB</div></div><div><br></div><div>now the node number 3 using 3x times more space than it should and node 1 and 2 2x times.</div>
<div><br></div><div>Also seems that after failure it is trying ot copy all data over the other node. It could be a good idea just to copy changed data. ( like drbd )</div><div>Is it any way to keep data in sync and delete the old data?</div>
<div><br></div>-- <br>--<br>Michael<br>
</div></div>