On Wed, Jun 3, 2015 at 4:29 PM, Joshua D. Drake <jd@commandprompt.com> wrote:
>
> On 06/03/2015 03:16 PM, Tomas Vondra wrote:
>
>> What is more important, though, is the amount of memory. OP reported the
>> query writes ~95GB of temp files (and dies because of full disk, so
>> there may be more). The on-disk format is usually more compact than the
>> in-memory representation - for example on-disk sort often needs 3x less
>> space than in-memory qsort. So we can assume the query needs >95GB of
>> data. Can you explain how that's going to fit into the 64GB RAM?
>>
>>> Cache is free memory. If you think of it any other way when you're
>>> looking at memory usage and pressure on theings like swap you're
>>> gonna make some bad decisions.
>>
>>
>> Cache is not free memory - it's there for a purpose and usually plays a
>> significant role in performance. Sure, it may be freed and used for
>> other purposes, but that has consequences - e.g. it impacts performance
>> of other queries etc. You generally don't want to do that on production.
>
>
> Exactly. If your cache is reduced your performance is reduced because less
> things are in cache. It is not free memory. Also the command "free" is not
> useful in this scenario. It is almost always better to use sar so you can
> see where the data points are that free is using.
But if that WAS happening he wouldn't still HAVE 60G of cache! That's
my whole point. He's NOT running out of memory. He's not even having
to dump cache right now.