Check for KVM Ballooning

klikli

Hello - how do I check if I am a victim of KVM ballooning/oversold RAM?

(Sorry for posting a screenshot here - copy-pasting results in a WAF block)

As you can see - the available memory is suspiciously low at 900M - despite the fact that there is not much memory consumption.

First-Root

Regarding the free: http://www.linuxatemyram.com

vovler

• dont buy anything from lowendbox
• pay more than the price of a coffee per month to a reputable provider
• get a dedi

klikli

FR_Michael said: Regarding the free: http://www.linuxatemyram.com

Right, I am aware of this.

It says, check the "available" column. It's exactly what I am pointing out - only 891.7MB available.

First-Root

@klikli said:

FR_Michael said: Regarding the free: http://www.linuxatemyram.com

Right, I am aware of this.

It says, check the "available" column. It's exactly what I am pointing out - only 891.7MB available.

Ballooning:
If you just want to see if "you are victim of ballooning" (weird sentence) you want to check the total column. That's the amount of ram your vps has "allocated" and is aware of at the moment. In case of ballooning this amount might be lower than the amount of ram you bought.

Oversold Ram:
you can't see that.

SpeedBus

Some of the memory is used by the kernel and is reserved -- this will vary depending on the kernel version in-use too.

You can use the command,

dmidecode -t memory

To list the RAM inserted into your VPS,

For example,

free -m output shows 1993 MB, which is around 55 MB lower than the 2G allocation,

              total        used        free      shared  buff/cache   available
Mem:           1993         293         434          18        1265        1498
Swap:          1021           0        1021

But the RAM inserted into the VPS is correct at 2 GB,

# dmidecode 3.1
Getting SMBIOS data from sysfs.
SMBIOS 2.4 present.

Handle 0x1000, DMI type 16, 15 bytes
Physical Memory Array
        Location: Other
        Use: System Memory
        Error Correction Type: Multi-bit ECC
        Maximum Capacity: 2 GB
        Error Information Handle: Not Provided
        Number Of Devices: 1

Handle 0x1100, DMI type 17, 21 bytes
Memory Device
        Array Handle: 0x1000
        Error Information Handle: 0x0000
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 2048 MB
        Form Factor: DIMM
        Set: None
        Locator: DIMM 0
        Bank Locator: Not Specified
        Type: RAM
        Type Detail: None

Next, you can check cat /proc/meminfo,

cat /proc/meminfo

which gives you the complete breakup of where each KB of RAM has gone

klikli

Thanks for the pointer to /proc/meminfo. Using the query here which sorts meminfo - looks like SUnreclaim is unusually large at 941 MB for whatever reason. Reboot fixed it.