Yes, I have looked at that picture. Do you really think you have only 25 processes running on that system? Sort your processes by Res and it will show the processes that are using the most of the RAM. If you sort by a criteria that is not related to memory the process you search can be #26 and will not show up.
If you don't even consider to try what is recommended to you, why are you asking in the first place?

We can't know that from the info posted.

The "virtualization" is hiding too much info.

The OP has said that is reasonable. I don't want to speculate whether that is right or wrong.

Do you know enough about this specific type of "virtualization" to justify that faith you're putting in that free RAM stat? I don't. Anything we know from non virtual Linux and true virtual Linux can be wrong when applied to this not quite virtual method of sharing a system.

As for using zero swap, that is incredibly unlikely, given all the other stats. I am pretty sure that the "virtualization" is simply stopping the tools from being able to see things like swap use.

Now you're giving good advice.

That info seems to be bouncing off the OP to no effect. I don't know a better way to phrase it.

If top was showing 25 processes and you then sort by RES, top will still be showing you only 25 processes, but it won't be the same 25 processes (merely sorted differently). It will be the 25 processes using the most memory.

I don't know how much of the memory use will show up that way. Maybe no more than we've already seen (but that's unlikely). Maybe very close to the total implied by free. Maybe significantly more than the total implied by free (even if the stats were real, there are many reasons any of those might happen).

I think this, not really virtual, shared system is showing you very distorted statistics for any system wide total (obviously not a true system wide total, but also not an accurate measure of your share).

But until you look at the understandable use of memory by processes, you're not ready to even start thinking about more complicated aspects.

Johnsfine -

As far as the physical host server is using physical RAM - we simply don't know.

But everything I've read seems to indicate the OP is worried about seeing 1++GB used in "top", and believing there might be some way to use less. It sounds like the OP is also concerned there might be some rogue process eating RAM.

Neither concern is warranted - "top" shows a perfectly happy system.

ALTERNATE CONCERN:
Maybe the underlying physical host is underpowered. Maybe the OP has control over the physical host and the virtualization software - maybe he doesn't.

But that's a separate question. As far as his Linux server: THERE IS NO PROBLEM!

PS:
The OP was also given a couple of "ps" commands to try, in addition to "top".

PPS:
That's in incorrect statement.

The "virtualization" hides ALL info. If you suspect the physical host, you must look at the physical host. You can't really tell ANYTHING about whether the physical host might be CPU-bound, I/O bound or memory-bound from inside the VM itself...

PPS:
The advice about JConsole (or equivalent) is also good advice. Whatever resources the VM is using, they appear to be Java-related.

Ok thankyou! I was just confused at why the pysical memory didn't add up

I am clearly just picking hairs trying to make my system run better!

Ok I think I understand now. As I am using zero swap, my pysical RAM usage is quite high. I am slowly getting closer to my RAM limit, and need to create a swapfile incase it crashes. How would I go about doing this, and will this decrease the performance of my server as a result?

Thanks!

Damn. After getting so far, i got to swapon: /swapfile1: Operation not permitted. Looks like my host has restricted it so I am stuck with zero swap!