August 12, 20178 yr I have more memory in use than I can find allocated to processes, I should have around 23/24GB in use out of 32GB (after starting the VM discussed below) Any suggestions how to track down the missing RAM usage? I've been running a second Windows VM with a GPU and today was unable to start it due to unable to allocate memory. It required 8GB, I have 32GB and my system showed 76% in use. Reducing the VM to 6GB allowed it to start taking me to 95% which grew to 96% a few minutes later. root@Tower:~# ps aux | awk '{print $6/1024 " MB\t\t" $11}' | sort -n [Cut around 40 processes all less than 100mb, most less than 30mb, everything above 100MB is shown below] 109.688 MB /opt/splunk/bin/python 116.438 MB /usr/local/sbin/shfs 124.473 MB node 137.062 MB mono 151.145 MB splunkd 176.051 MB java 193.863 MB bin/mongod 205.191 MB /usr/lib/jvm/java-7-openjdk-amd64/bin/java 271.27 MB mono 321.172 MB /usr/bin/java 394.312 MB mono 1587.43 MB java 4113.69 MB /usr/bin/tvheadend 4332.82 MB /usr/bin/qemu-system-x86_64 6226.94 MB /usr/bin/qemu-system-x86_64 8331.22 MB /usr/bin/qemu-system-x86_64
August 12, 20178 yr VM's require a block of RAM to manage each instance, for the emulated hardware and stuff. I'd allocate at least 1 or 2GB per running VM so the host OS has enough spare to keep things running smoothly.
August 12, 20178 yr Also.. Linux always utilises all available RAM... So you mostly have nothing left no matter how much you are using..
August 12, 20178 yr Author I have three VM's running and IMO should have 8GB free. Something is using a lot more than it should and I can't seem to see what. root@Tower:~# free -m total used free shared buff/cache available Mem: 32128 29076 350 871 2701 1526 Swap: 0 0 0 root@Tower:~#
August 12, 20178 yr Author 2 minutes ago, Helmonder said: Also.. Linux always utilises all available RAM... So you mostly have nothing left no matter how much you are using.. But it makes it available when requested and in this case the VM wouldn't start. Something is holding onto it. EDIT - I was running the VM with 8GB ram for the last two nights, nothing has changed but the VM today will only start with 6GB. I understand I have a lot running and things change but I'm just trying to see what is using it. I expect if i reboot I'll have both VM's running with 8GB along with everything else. Edited August 12, 20178 yr by Tuftuf
August 12, 20178 yr 4 hours ago, Tuftuf said: I was running the VM with 8GB ram for the last two nights, nothing has changed but the VM today will only start with 6GB. Memory fragmentation. It has to be able to allocate the entire block contiguously, if there is a process that hasn't release RAM it can keep a block that size from being available.
August 12, 20178 yr Author 33 minutes ago, jonathanm said: Memory fragmentation. It has to be able to allocate the entire block contiguously, if there is a process that hasn't release RAM it can keep a block that size from being available. 1 Is that based on an understanding of Unraid or KVM? It's the complete opposite to the RHEL docs for running KVM for example. I still don't agree it's simply memory fragmentation. Unraid generally shows me a % of how much memory is in use, I don't know for certain where this value is taken from but this value generally gives a good indication to how much is in use by VM's for example (also other stuff but they are the ram heavy resource). I've run many VM's on this system already, right now i'm trying to hunt down what is causing the memory % to be so high when I can't find it in processes/memory usage. I'm not saying this is a bug but I wanted to find what was using the memory. I'll reboot and see! EDIT - root@Tower:~# free -m total used free shared buff/cache available Mem: 32128 22494 333 573 9300 8417 Swap: 0 0 0 root@Tower:~# TVHeadend and Java were using a lot less after a reboot. I can control TVH's usage but will need to find out what was using java. nevermind! Edited August 12, 20178 yr by Tuftuf
Archived
This topic is now archived and is closed to further replies.