<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Thu, Sep 18, 2014 at 7:33 AM, Gavin Panella <span dir="ltr"><<a href="mailto:gavin.panella@canonical.com" target="_blank">gavin.panella@canonical.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">...<br>
<span class="">>> For DEPLOYED nodes, sure, the command will currently be lost, but<br>
>> these nodes are, one assumes, under active management, and some<br>
>> process outside of MAAS will notice, be that a human or a Juju or<br>
>> something else.<br>
><br>
> I think that's a dreadful user experience. We should not be knowingly<br>
> throwing away user requests.<br>
<br>
</span>It's not perfect, but it's far from dreadful.<br>
<br>
MAAS can never provide a perfect service. Even if we coded everything<br>
perfectly, it would still suffer from failures outside of its control,<br>
like hardware failure, poor quality BMC firmware, human error, fire,<br>
flood, war, and so on. Not resuming power commands after a crash or<br>
restart of a cluster is a failure mode that we can iterate on and<br>
reduce, but it's not the end of days.<br></blockquote><div><br></div><div>Throwing away user requests is very much dreadful. Because of exactly the reasons you mentioned above - there are a lot of things outside of MAAS that can impact what MAAS is trying to manage - conistency and accuracy is key. If I in fact can't do something, then you error and tell me, you don't drop the request on the floor.</div><div><br></div><div>Thanks,</div><div>Dean</div></div></div></div>