<div dir="ltr"><br><div class="gmail_extra"><br><br><div class="gmail_quote">2014-08-18 11:11 GMT+02:00 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"><div class="">
> Restore: 6'01''<br>
<br>
</div>We use only at most nr_of_nodes * 2 threads for snapshot save, considering not<br>
draining the cluster too much.<br>
<br></blockquote><div>Ok clear.<br> <br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I guess qemu-img use more threads and it doesn't calculating the fingerprint<br>
to de-duplicate.<br></blockquote><div><br></div><div>Repeating for the second time 'cluster snapshot save' with '-m' option, it takes 3'37''.<br></div><div>I guess because fingerprints were already calculated. <br>
<br></div><div>Import is also a bit slower:<br>5'05'' for 'cluster snaphsot load -m' <br>3'43'' for qemu-img.<br><br></div><div>In case of selective vdi restore it's fine like that, to avoid stealing too much resources as you say.<br>
</div></div></div></div>