<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<br>
<br>
<div class="moz-cite-prefix">On 02/05/2016 04:26 PM, Gilberto Nunes
wrote:<br>
</div>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>> Also if you have no more VM under HA you can remove
the HA resource config<br>
><br>
> # rm /etc/pve/ha/resources.cfg<br>
<br>
</div>
But if do this, in the future, when I realize that I need insert
the previously node removed, it will work?<br>
</div>
</blockquote>
<br>
Yes, if you add the other node again and want to reactivate HA
simply add a VM/CT to HA (via GUI or the command line) and this file
will be recreated if not existing and everything will work again.<br>
<br>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div> <br>
<br>
> But the errors from the CRM should _not_ be related to a
stopping VM, I'm pretty sure there's another > cause for
that :)<br>
<br>
</div>
<div>That's is what I am trying figure out.<br>
<br>
</div>
<div> <br>
> If you have the logs from around the time the VM stopped
look if something strange happened. <br>
<br>
</div>
<div>Well... This happen later of night and, seems to me that is
happen when backup go...<br>
</div>
</div>
</blockquote>
OOM-killer because you get out of RAM?<br>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>I notice, in other PVE host, sometimes some randonly VM's
just restart at night...<br>
</div>
</div>
</blockquote>
To be honest, that seems like it comes from inside the VM not from
PVE.<br>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>I any case, logs from proxmox are where difficult to get,
mainly because the same of syslog system and I get a lot of
garbage...<br>
</div>
</div>
</blockquote>
I really can't relate to that i find the logs quite practical to
skim through, especially with the journalctl tool.<br>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>We have the /var/log/pve and /var/log;pveproxy/ but nothing
useful at all....<br>
</div>
</div>
</blockquote>
This is also not the primary log output.<br>
<blockquote
cite="mid:CAOKSTBsPHAnz5KKtX20Fp4y=w4A9n8LdqwoY7T3Ab39nX9eHzQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div><br>
</div>
<div>I will try another approuch to solve this issues....<br>
<br>
</div>
<div>Thanks anyway....<br>
</div>
</div>
</blockquote>
<br>
btw. you can pack a lot of relevant things with journalctl, it has
some nifty functions, e.g.: you can create a log from a date/time
frame:<br>
<blockquote type="cite">
<pre class="code-pre "><code langs="">journalctl --since "2015-01-10" --until "2015-01-11 03:00" </code></pre>
</blockquote>
<br>
(additional <br>
<pre class="code-pre "><blockquote type="cite"><pre class="code-pre "><code langs="">> out.log </code>
</pre></blockquote></pre>
at the end, to pipe it in a file.)<br>
<br>
<br>
</body>
</html>