New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Memory useage problem?
Posted this over on the quality servers 128meg thread as this is one of their boxes, but thought id post here as well for a wider audience!
Ok, so what’s going on with memory usage? Is it me not being able to read/add up or do i need to open a ticket? This is Debian 6 32bit after running lowendscript ‘system’ part only to clean up, then a fresh reboot.
root@QLS128:~# ps aux
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.5 2028 680 ? Ss 16:29 0:00 init [2]
root 307 0.0 0.5 1948 688 ? S 16:29 0:00 /usr/sbin/syslo
root 339 0.0 0.5 2288 764 ? Ss 16:29 0:00 /usr/sbin/cron
root 354 0.0 0.6 2392 908 ? Ss 16:29 0:00 /usr/sbin/xinet
root 557 1.0 0.8 2400 1116 ? Ss 22:13 0:00 dropbear -i
root 558 0.0 1.2 2960 1612 pts/0 Ss 22:13 0:00 -bash
root 562 0.0 0.6 2348 908 pts/0 R+ 22:13 0:00 ps aux
root@QLS128:~# free
total used free shared buffers cached
Mem: 131072 32020 99052 0 0 6500
-/+ buffers/cache: 25520 105552
Swap: 0 0 0
root@QLS128:~#
To me, this looks like the processes are using a fraction of what ‘free’ is reporting as used. On other VM’s I have, these figures are far closer.
What goes on? Anyone have any idea?
Comments
Since you have a VPS running on OpenVZ have you check the vzFree script from LEB.
You can find more details here
the vzfree script is a little outdated on newer openvz versions. Make sure you are looking at the -/+ buffers/cache line.
In this exact example your VZ instance is caching a lot of your RAM for file system use. You can (At your own risk) adjust these.. Well maybe not -- I don't know how restrictive it would be in a container ... Google it
Thanks for the feedback guys, as long as there's not something inherently wrong that's ok. Is any OS going to cache less or is it down to how the container is set up?
I was worrying a bit as its so different to my 128meg BuyVM.
It's different because it's running a newer version of OpenVZ. The caches will be cleared to make room for other things that need RAM.
Seeing you have 0 swap on your free output.
QualityServers uses VSwap, but they have a problem with configuring the VSwap when you install Debian 6.0 32bit on the VPS. The LowEndBox offer was with 128MB VSwap, and you should have 128MB of swap in the free output. There is a thread about this in the QualityServers forum.
Thankd for the heads up Leo, will check it out.
Posted on the thread and James had fixed it 10 minutes later, Cant knock that!
So what was the issue? The missing vswap?
That's what James fixed yeah, but I hadn't noticed the vswap was missing till Leo pointed it out.
Originally I was concerned that I seemed to be missing memory when comparing what my processes were using with what 'free' was reporting, but i wasn't taking buffers/cache into account. I'm still learning a lot about linux
I think mainly i was worried as i was seeing such a difference between my qualityservers 128meg box and my BuyVM 128meg box, when I thought they should be pretty much the same. Sounds like its down to different versions and configurations on OpenVZ.
The server is running well, using it to host a teamspeak 3 server for about 20 gaming mates.