Thanks for your help.I now wonder how to simulate a node goes down.Does kill -9 work?<br><br>Best.<br><br><div class="gmail_quote">2012/9/13 Liu Yuan <span dir="ltr"><<a href="mailto:namei.unix@gmail.com" target="_blank">namei.unix@gmail.com</a>></span><br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On 09/13/2012 11:49 AM, Tao Lin wrote:<br>
> �I'm new� to sheepdog, and I deployed a sheepdog cluster with 3 nodes,<br>
<div><div class="h5">> using zookeeper.So far, everything seems work well,I'm now looking for<br>
> the correct way of restoring cluster if a certain node goes down.Is it<br>
> just like starting a new node? e.g. sheep -d /media/store4sheep -c<br>
> zookeeper:192.168.99.1.<br>
<br>
</div></div>Yes, you can simply hot plug in the failed node back by starting sheep<br>
daemon. Nothing else you need to do. Data will be automatically<br>
re-balanced and recovered.<br>
<br>
You can easily scale out the cluster by hot plug in more nodes by<br>
starting sheep daemon on the new nodes. You can hot unplug the nodes<br>
safely by collie if you want to shrink the storage.<br>
<br>
Thanks,<br>
Yuan<br>
</blockquote></div><br>