Changes

{{UBC tocWarning|The information below is obsoleted, you should just use [[VSwap]].}}

=== Question: I want to know how much RAM is allocated to a particular VEcontainer. I ran cat /proc/user_beancounters from inside my VPS container and I have the following results below. But I am not sure how to interpret the results. ===

The guarantee is vmguarpages barrier = 125000 pages. In any case, the VPS container will be able to allocate that size of memory. Well... in almost any case, with one exception...

...when the whole node is in out of memory situation. In that case, the guarantee is oomguarpages barrier (the same 125000 pages in this example). It means that no VPS container process will be killed if memory usage by the VPS container is less than 125000 * 4096 bytes. If it is more, some process can be killed, and failcounter of oomguarpages will be increased.

vmguarpages barrier is a guarantee: in any case (except out-of-memory aka OOM) you will be able to allocate 125000 pages. Well, probably you will be able to allocate more, up to 750000 pages (and if you are cool high priority process, you will be able to allocate 825000 pages, but definitely not more). We don't say anything about RAM or SWAP here. These pages can be swapped by HWnode kernel if needed, and can live in RAM if possible. Frankly, you can't say "I want this VPS container to have X MB RAM and Y MB SWAP". There is no such parameter.

physpages parameter accounts total amount of RAM used by the VPS container (the memory shared between VPSes containeres is divided and accounted in physpages for all VPSes containeres which use it in equal fractions). It cannot be limited, it just shows the current situation.

the sum is 247869844 bytes. If this sum is more than oomguarpages barrier = 125000 pages = 512000000 bytes (it is not), some of the processes on the VPS container can be killed in case of OOM situation.

=== Question: Is there a way to know for a VPS container account holder (me) that a server has reached an OOM situation? ==='free' inside a VPS container will show memory statistics for the hardware node (total memory, used memory, total swap, used swap). It's OK for Linux system to use as much RAM as possible, so free memory size is going to be not more than a few megabytes.

<pre>Rojoblandino wrote:Thank you for everything, unfortunately the only way I have to help is to work with code,I leave something here that I hope you will enjoy and see it very usefull. And yes therewas not any script but now there is one. Greetings and have fun.</pre> <pre>#!/bin/bash# Openvz Containers Memory Status# Version 1.5# Reference: http://wiki.openvz.org/Setting_UBC_parameters # The program is placed under the GNU Public License (GPL)# Author: Roberto Blandino Cisneros.# Under any change please notify me# rojoblandino(at)yahoo.com # Advantages or changes:# 1) Calculation of vz containers.# 2) Percentage Calculation.# 3) Avoid simples wrong characters and out of range entry.# 4) Use the same same order of vzlist output.# 5) Problems on vzid param mistakes fixed.# 6) Fixed when 0 passed by param an error appear # The usage of this script is easy.# To understand how it works first you must know# the quantity of your containers so it is the# first thing we do.# Each container get a number if you have three# containers e.g. eva gira huna# Depending of the order ar shown in # /proc/user_beancounters that number of position# will be taken.# If eva, gira and huna containers are shown in that# order they will be 1 2 and 3 respectively.# Is so simple like that, then if you want to see# the status for gira container you must type in# your shell:# $> bash vmstatus.sh 2# And you will see the status of this container. # First thing to do is to get the list of containersvzn=$(($(vzlist | wc -l)-1)) case $# in 1) # Get the container to be calculated vm_param=$1  # Just avoiding our human error datachk=$(echo $vm_param | grep -e [a-z] | wc -l) if let "datachk>0";then printf "That's not a valid number!\n" exit fi  # Avoiding another mistake maybe the user forgot the quantity of vm allocated if let "vm_param>vzn";then printf "Container out of range, please try a number between 1 and %d\n" "$vzn" exit elif let "vm_param==0";then printf "This value is not available, value must be higher than 0 \n" exit fi  # Take the real position of the container to get from the vzlist # for avoiding the first line we add one line more, this mean that id # will be taken from the 2nd line ahead let $vm "vm=vm_param+1"  # Uncomment this and comment the line above if you want to switch the order # let $vm "vm=(n-vm_param)+2"  # Take the id from the vzlist output id_vm=$(vzlist | head -$vm | tail -1|awk '{print $1}')  # Get the line where is located the vm to be calculated val=$(grep -n $id_vm /proc/user_beancounters | awk -F: '{print $1}')  # Adding the 23 block where the info is located let $val "val=23+(val)"  # If you want to see all values of the vm uncomment this #head -$val /proc/user_beancounters | tail -24  # The math start here. # To understand what i am doing here i will repeat again that # you must read carefully the page http://wiki.openvz.org/Setting_UBC_parameters # the only thing i do was to get the values explained already there. kmemsize=$(head -$val /proc/user_beancounters | tail -24 | head -1 | tail -1 | awk '{print $3}') oomguarpages=$(head -$val /proc/user_beancounters | tail -24 | head -9 | tail -1 | awk '{print $2}') tcpsndbuf=$(head -$val /proc/user_beancounters | tail -24 | head -14 | tail -1 | awk '{print $2}') tcprcvbuf=$(head -$val /proc/user_beancounters | tail -24 | head -15 | tail -1 | awk '{print $2}') othersockbuf=$(head -$val /proc/user_beancounters | tail -24 | head -16 | tail -1 | awk '{print $2}') dgramrcvbuf=$(head -$val /proc/user_beancounters | tail -24 | head -17 | tail -1 | awk '{print $2}') oomguarpages_barrier=$(head -$val /proc/user_beancounters | tail -24 | head -9 | tail -1 | awk '{print $4}')  # Three current values need to be calculated: # oomguarpages current value (MEM+SWAP actual usage) # socket buffers current value # kmemsize current value let $resp "resp=((oomguarpages*4096)+tcpsndbuf+tcprcvbuf+othersockbuf+dgramrcvbuf+kmemsize)"  # Calculating the oom Barrier let $oom "oom=(oomguarpages_barrier*4096)"  # Getting the percentage of the output let $porc "porc=(resp*100)/oom"  # Showing the ID and Percentage of the container #echo $(head -$val /proc/user_beancounters | tail -24| head -1 | awk -F: '{print $1}')" Status: "$porc"%" printf "%s Status: %d %%\n" "$(head -$val /proc/user_beancounters | tail -24| head -1 | awk -F: '{print $1}' | sed "s/^\ *//g")" "$porc"  # Showing the two important values. string1="System Count:" string2="OOM Barrier:" ls1=${#string1} ls2=${#resp} if let "ls1<ls2";then let $ls3 "ls3=ls2-ls1" for (( i=0;i<$ls3;i++ ));do space_line=$space_line" " done printf "System Count:%s\t-\tOOM Barrier:\n" "$space_line" printf "%s\t-\t%s\n" "$resp" "$oom" elif let "ls1>ls2";then let $ls3 "ls3=ls1-ls2" for (( i=0;i<$ls3;i++ ));do space_line=$space_line" " done printf "System Count:\t-\tOOM Barrier:\n" printf "%s%s\t-\t%s\n" "$resp" "$space_line" "$oom" else printf "System Count:\t-\tOOM Barrier:\n" printf "%s\t-\t%s\n" "$resp" "$oom" fi # Uncoment this if you just want to now if it is below to barrier # it means is healthy if resp is less than oom # if let "resp<oom";then # Or let the following if you want to evaluate as a percentage # 75% is by default you must change it to your own warning barrier. if let "porc<75";then echo "It is ok!" else echo "Some of the processes will be killed!" fi ;; *) echo "Just one parameter is allowed for this script" echo "e.g. "$0" 1 " ;; esac</pre> === Can you please elaborate the difference between oomguarpages and privvmpages?=== Both seem to be showing mem + swap. Is the difference between two about allocated but not used pages? I mean, privvmpages show allocated (some of which might not be used) for mem + swap, whereas oomguarpages show allocated, all of which are used, mem + swap? ====