<div dir="ltr"><div class="gmail_extra">On Fri, Nov 1, 2013 at 6:39 PM, Dietmar Maurer <span dir="ltr"><<a href="mailto:dietmar@proxmox.com" target="_blank">dietmar@proxmox.com</a>></span> wrote:<br><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div class="im">> If one of this actions fail, it MUST be restored on current node.<br>
<br>
</div>At some point it is no longer possible to restore on current node.<br></blockquote><div><br></div><div>At what point? </div><div>Ok, look at online migration.</div><div>We can restore container at any time. After 1st sync(actually vm not stopped, and we can only cancel sync)</div>
<div>After suspend(we get vm dump-file and we can restore it) if any of following actions will fail:</div><div>2nd sync, copying of dump-file<br></div><div>And finally - ct restore.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div class="im"><br>
> Now we can see something like this:<br>
><br>
> Nov 01 13:32:55 container is running - using online migration Nov 01 13:32:56<br>
> starting rsync phase 1<br>
> -----------<br>
> Nov 01 13:34:18 start live migration - suspending container Nov 01 13:34:18<br>
> dump container state Nov 01 13:34:18 copy dump file to target node Nov 01<br>
> 13:34:20 starting rsync (2nd pass) Nov 01 13:34:20 dump 2nd level quota Nov 01<br>
> 13:34:20 copy 2nd level quota to target node Nov 01 13:34:22 initialize<br>
> container on remote node 'node2'<br>
> Nov 01 13:34:22 initializing remote quota Nov 01 13:34:22 turn on remote quota<br>
> Nov 01 13:34:22 load 2nd level quota Nov 01 13:34:22 starting container on<br>
> remote node 'node2'<br>
> Nov 01 13:34:22 restore container state<br>
> Nov 01 13:34:23 Restoring container ...<br>
> Nov 01 13:34:23 Starting container ...<br>
> Nov 01 13:34:23 Container is mounted<br>
> Nov 01 13:34:23 undump...<br>
> Nov 01 13:34:23 Setting CPU units: 1000<br>
> Nov 01 13:34:23 Setting CPUs: 1<br>
> Nov 01 13:34:23 Setting devices<br>
> Nov 01 13:34:23 Error: undump failed: Cannot allocate memory Nov 01 13:34:23<br>
<br>
</div>You just do not have enough free memory on that node?<br>
<br>
</blockquote></div><div class="gmail_extra"><br></div>Yes, there is 64G free.</div><div class="gmail_extra">Same problem, when vm disk space is 100% use(simfs). There is a error in container start <br>No matter, what error i get. If there is an error, container must be restored on first node with condition, that it has on suspend.</div>
<div class="gmail_extra">vzmigrate(openvz package) do it well.</div></div>