I've observed high memory usage at amazon ec2 instance using amazon Linux.

See attached image. enter image description here Top 25 processes shows < 1GB of memory, but free -m says ~ 3GB is consumed. I've already visited http://www.linuxatemyram.com/, but in my case cache is also displayed as consumed.

EDIT: I've low free count in -/+ buffers/cache in 2nd line of output.

Any idea?

Output from cat /proc/meminfo

MemTotal:        3857992 kB
MemFree:           85032 kB
MemAvailable:    3589668 kB
Buffers:          102704 kB
Cached:           176500 kB
SwapCached:         1876 kB
Active:           358416 kB
Inactive:          66924 kB
Active(anon):     159468 kB
Inactive(anon):    27128 kB
Active(file):     198948 kB
Inactive(file):    39796 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       8388604 kB
SwapFree:        8377616 kB
Dirty:                24 kB
Writeback:             0 kB
AnonPages:        144268 kB
Mapped:            45708 kB
Shmem:             40460 kB
Slab:            3308524 kB
SReclaimable:    3295052 kB
SUnreclaim:        13472 kB
KernelStack:        1744 kB
PageTables:         9056 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:    10317600 kB
Committed_AS:     491944 kB
VmallocTotal:   34359738367 kB
VmallocUsed:       11340 kB
VmallocChunk:   34359722271 kB
AnonHugePages:         0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:       22528 kB
DirectMap2M:     3909632 kB
  • Possible duplicate of Meaning of the buffers/cache line in the output of free – yoonix Jan 27 '16 at 9:53
  • 1
    Probably NO, because I actually have low free in 2nd line, according to your link: "A Linux system is really low on memory if the free value in -/+ buffers/cache: line gets low." – Piyuesh Jan 27 '16 at 10:28
  • 1
    what does cat /proc/meminfo say? – user1700494 Jan 27 '16 at 10:36
  • 613MB is not low especially considering your swap isn't being used hard. The 3154MB in buffers/cache should be subtracted from the Mem used number. 3421-3154=267MB used by something other than the cache. That 267MB should be pretty close to what you see in ps or top. – yoonix Jan 27 '16 at 10:49
  • try checking your memory with free / top before and after a: (as root) echo 3 > /proc/sys/vm/drop_caches, you should see a big difference. Check here for more info on that: kernel.org/doc/Documentation/sysctl/vm.txt (kernel documentation) – Fredi Jan 27 '16 at 11:03

SOLVED: As per http://blog.famzah.net/2014/09/24/memavailable-metric-for-linux-kernels-before-3-14-in-procmeminfo/ actual free memory is determined by MemAvailable in /proc/meminfo output.

Actual issue here explained at Unusually high dentry cache usage

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.