<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
</head>
<body>
<div style="color: black;">
<p style="margin: 0 0 1em 0; color: black; font-family: sans-serif;">On 7 November 2017 10:33:46 am Mark Perkins <perkins1724@hotmail.com> wrote:</p>
<p style="margin: 0 0 1em 0; color: black; font-family: sans-serif;">> On 7 November 2017 9:58:43 am Ian Evans <dheianevans@gmail.com> wrote:<br>
><br>
>>> As was already mentioned that swapping should be checked. I don't know<br>
>>> what your top showed from 0.27 times but 20% CPU load for three recordings<br>
>>> plus a commflag job on 2011 hardware sounded fine to me. I couldn't see the<br>
>>> problem there. But swapping - particularly 500+ MiB - that sounds awful and<br>
>>> sounds like the symptoms you are seeing.<br>
>>><br>
>>> Try shift + m while in top to sort by % memory used and post that.<br>
>>><br>
>>> Do you accidentally have multiple FE processes running?<br>
>>><br>
>>><br>
>> Of course when I NEED a spike I don't get one. Here's a top with one<br>
>> recording and commflag. Load is over 1 though. Will try again if I see a 20+<br>
>><br>
>><br>
><br>
> I suspect that the high load will return as soon as your system starts swapping lots of mem again. But I think the issue (cause) is probably the mem swapping.<br>
><br>
><br>
><br>
><br>
> ----------<br>
> _______________________________________________</p>
<p style="margin: 0 0 1em 0; color: black;"></p>
<p style="margin: 0 0 1em 0; color: black;">Specifically this might be an interesting read -
<a href="http://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html">http://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html</a></p>
<p style="margin: 0 0 1em 0; color: black;">Linux load averages also reflect demand for other system resources. Including swap usage.
</p>
<p style="margin: 0 0 1em 0; color: black;">For CPU utilisation using the per-process figures from top might be better.
</p>
</div>
</body>
</html>