25

August 2015 Summary

Please note, this is still happening. This is not related to linuxatemyram.com - the memory is not used for disk cache/buffers. This is what it looks like in NewRelic - the system leaks all the memory, uses up all swap space and then crashes. In this screenshot I rebooted the server before it crashed:

enter image description here

It is impossible to identify the source of the leak using common userspace tools. There is now a chat room to discuss this issue: http://chat.stackexchange.com/rooms/27309/invisible-memory-leak-on-linux

Only way to recover the "missing" memory appears to be rebooting the server. This has been a long standing issue reproduced in Ubuntu Server 14.04, 14.10 and 15.04.

Top

The memory use does not show in top and cannot be recovered even after killing just about every process (excluding things like kernel processes and ssh). Look at the "cached Mem", "buffers" and "free" fields in top, they are not using up the memory, the memory used is "missing" and unrecoverable without a reboot.

Attempting to use this "missing" memory causes the server to swap, slow to a crawl and eventually freeze.

root@XanBox:~# top -o +%MEM
top - 12:12:13 up 15 days, 20:39,  3 users,  load average: 0.00, 0.06, 0.77
Tasks: 126 total,   1 running, 125 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.1 us,  0.2 sy,  0.0 ni, 99.7 id,  0.0 wa,  0.1 hi,  0.0 si,  0.0 st
KiB Mem:   2,040,256 total,  1,881,228 used,    159,028 free,     1,348 buffers
KiB Swap:  1,999,868 total,     27,436 used,  1,972,432 free.    67,228 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
11502 root      20   0  107692   4252   3240 S   0.0  0.2   0:00.06 sshd: deployer [priv]
11336 root      20   0  107692   4248   3240 S   0.0  0.2   0:00.06 sshd: deployer [priv]
11841 root      20   0  107692   4248   3240 S   0.0  0.2   0:00.06 sshd: deployer [priv]
11301 root      20   0   26772   3436   2688 S   0.7  0.2   0:01.30 /usr/sbin/openvpn --writepid /var/run/openvpn.zanview.com.pid --status /var/run/openvpn.zanview.com.status 10 --cd /etc/openvpn --config /etc/openvpn/z+
11385 deployer  20   0   19972   2392   1708 S   0.0  0.1   0:00.03 -bash
11553 deployer  20   0   19972   2388   1708 S   0.0  0.1   0:00.03 -bash
11890 deployer  20   0   19972   2388   1708 S   0.0  0.1   0:00.02 -bash
11889 deployer  20   0  108008   2280    944 S   0.0  0.1   0:00.25 sshd: deployer@pts/3
12009 root      20   0   18308   2228   1608 S   0.0  0.1   0:00.09 -su
12114 root      20   0   18308   2192   1564 S   0.0  0.1   0:00.04 -su
12007 root      20   0   67796   2136   1644 S   0.0  0.1   0:00.01 sudo su -
12112 root      20   0   67796   2136   1644 S   0.0  0.1   0:00.01 sudo su -
12008 root      20   0   67376   2016   1528 S   0.0  0.1   0:00.01 su -
12113 root      20   0   67376   2012   1528 S   0.0  0.1   0:00.01 su -
    1 root      20   0   33644   1988    764 S   0.0  0.1   2:29.77 /sbin/init
11552 deployer  20   0  107692   1952    936 S   0.0  0.1   0:00.07 sshd: deployer@pts/2
11384 deployer  20   0  107692   1948    936 S   0.0  0.1   0:00.06 sshd: deployer@pts/0
12182 root      20   0   20012   1516   1012 R   0.7  0.1   0:00.08 top -o +%MEM
 1152 message+  20   0   39508   1448    920 S   0.0  0.1   1:40.01 dbus-daemon --system --fork
 1791 root      20   0  279832   1312    816 S   0.0  0.1   1:16.18 /usr/lib/policykit-1/polkitd --no-debug
 1186 root      20   0   43736    984    796 S   0.0  0.0   1:13.07 /lib/systemd/systemd-logind
 1212 syslog    20   0  256228    688    184 S   0.0  0.0   1:41.29 rsyslogd
 5077 root      20   0   25324    648    520 S   0.0  0.0   0:34.35 /usr/sbin/hostapd -B -P /var/run/hostapd.pid /etc/hostapd/hostapd.conf
  336 root      20   0   19476    512    376 S   0.0  0.0   0:07.40 upstart-udev-bridge --daemon
  342 root      20   0   51228    468    344 S   0.0  0.0   0:00.85 /lib/systemd/systemd-udevd --daemon
 1097 root      20   0   15276    364    256 S   0.0  0.0   0:06.39 upstart-file-bridge --daemon
 4921 root      20   0   61364    364    240 S   0.0  0.0   0:00.05 /usr/sbin/sshd -D
  745 root      20   0   15364    252    180 S   0.0  0.0   0:06.51 upstart-socket-bridge --daemon
 4947 root      20   0   23656    168    100 S   0.0  0.0   0:14.70 cron
11290 daemon    20   0   19140    164      0 S   0.0  0.0   0:00.00 atd
  850 root      20   0   23420     80     16 S   0.0  0.0   0:11.00 rpcbind
  872 statd     20   0   21544      8      4 S   0.0  0.0   0:00.00 rpc.statd -L
 4880 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty4
 4883 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty5
 4890 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty2
 4891 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty3
 4894 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty6
 4919 root      20   0    4368      4      0 S   0.0  0.0   0:00.00 acpid -c /etc/acpi/events -s /var/run/acpid.socket
 5224 root      20   0   24048      4      0 S   0.0  0.0   0:00.00 /usr/sbin/rpc.mountd --manage-gids
 6160 root      20   0   14540      4      0 S   0.0  0.0   0:00.00 /sbin/getty -8 38400 tty1
    2 root      20   0       0      0      0 S   0.0  0.0   0:03.44 [kthreadd]
    3 root      20   0       0      0      0 S   0.0  0.0   1:04.63 [ksoftirqd/0]
    5 root       0 -20       0      0      0 S   0.0  0.0   0:00.00 [kworker/0:0H]
    7 root      20   0       0      0      0 S   0.0  0.0  16:03.32 [rcu_sched]
    8 root      20   0       0      0      0 S   0.0  0.0   4:08.79 [rcuos/0]
    9 root      20   0       0      0      0 S   0.0  0.0   4:10.42 [rcuos/1]
   10 root      20   0       0      0      0 S   0.0  0.0   4:30.71 [rcuos/2]

Hardware

I have observed this on 3 servers out of around 100 so far (though others may be affected). One is an Intel Atom D525 @1.8ghz and the other 2 are Core2Duo E4600 and Q6600. One is using a JMicron Technology Corp. JMC250 PCI Express Gigabit Ethernet Controller, the others are using Qualcomm Atheros Attansic L1 Gigabit Ethernet (rev b0).

I ran lshw on the trouble servers as well as on an example OK server. Problem Servers: http://pastie.org/10370534 http://pastie.org/10370537 and http://pastie.org/10370541 -- OK Server: http://pastie.org/10370544

Application

This is an entirely headless application. There is no monitor connected and in fact no XServer installed at all. This should rule out graphics drivers/issues.

The server is used to proxy and analyse RTSP video using live555ProxyServer, ffmpeg and openCV. These servers do crunch through a lot of traffic because this is a CCTV application: http://pastie.org/9558324

I have tried both very old and latest trunk versions of live555, ffmpeg and openCV without change. I have also tried using opencv through the python2 and python3 modules, no change.

The exact same software/configuration has been loaded onto close to 100 servers, so far 3 are confirmed to leak memory. The servers slowly and stealthily leak around xMB (one leaking 8MB, one is slower, one is faster) per hour until all ram is gone, the servers start swapping heavily, slow to a crawl and require a reboot.

Meminfo

Again, you can see the Cached and Buffers not using up much memory at all. HugePages are also disabled so this is not the culprit.

root@XanBox:~# cat /proc/meminfo
MemTotal:          2,040,256 kB
MemFree:             159,004 kB
Buffers:               1,348 kB
Cached:               67,228 kB
SwapCached:            9,940 kB
Active:               10,788 kB
Inactive:             81,120 kB
Active(anon):          1,900 kB
Inactive(anon):       21,512 kB
Active(file):          8,888 kB
Inactive(file):       59,608 kB
Unevictable:               0 kB
Mlocked:                   0 kB
SwapTotal:         1,999,868 kB
SwapFree:          1,972,432 kB
Dirty:                     0 kB
Writeback:                 0 kB
AnonPages:            14,496 kB
Mapped:                8,160 kB
Shmem:                    80 kB
Slab:                 33,472 kB
SReclaimable:         17,660 kB
SUnreclaim:           15,812 kB
KernelStack:           1,064 kB
PageTables:            3,992 kB
NFS_Unstable:              0 kB
Bounce:                    0 kB
WritebackTmp:              0 kB
CommitLimit:       3,019,996 kB
Committed_AS:         94,520 kB
VmallocTotal: 34,359,738,367 kB
VmallocUsed:         535,936 kB
VmallocChunk: 34,359,147,772 kB
HardwareCorrupted:         0 kB
AnonHugePages:             0 kB
HugePages_Total:           0
HugePages_Free:            0
HugePages_Rsvd:            0
HugePages_Surp:            0
Hugepagesize:          2,048 kB
DirectMap4k:          62,144 kB
DirectMap2M:       2,025,472 kB

Free Output

Free shows the following (note cached and buffers are both low so this is not disk cache or buffers!) - the memory is not recoverable without a reboot:

root@XanBox:~# free -m
             total       used       free     shared    buffers     cached
Mem:         1,992      1,838        153          0          1         66

If we subtract/add the buffers/cache to Used and Free, we see:

  • 1,772MB Really Used (- Buffers/Cache) = 1,838MB used - 1MB buffers - 66MB cache
  • 220MB Really Free (+ Buffers/Cache) = 154MB free + 1MB buffers + 66MB cache

Exactly as we expect:

-/+ buffers/cache:      1,772        220

So around 1.7GB is not used by userspace and in fact used by the kernel as the system is actually using 53.7MB (see PS Mem output below).

I'm surprised with the amount of comments that think 1.7GB is used for caching/buffers - this is fundamentally misreading the output! - this line means used memory excluding buffers/cache, see linuxatemyram.com for details.

PS Output

Here is a full list of running processes sorted by memory:

# ps -e -o pid,vsz,comm= | sort -n -k 2
    2      0 kthreadd
    3      0 ksoftirqd/0
    5      0 kworker/0:0H
    7      0 rcu_sched
    8      0 rcuos/0
    9      0 rcuos/1
   10      0 rcuos/2
   11      0 rcuos/3
   12      0 rcu_bh
   13      0 rcuob/0
   14      0 rcuob/1
   15      0 rcuob/2
   16      0 rcuob/3
   17      0 migration/0
   18      0 watchdog/0
   19      0 watchdog/1
   20      0 migration/1
   21      0 ksoftirqd/1
   23      0 kworker/1:0H
   24      0 watchdog/2
   25      0 migration/2
   26      0 ksoftirqd/2
   28      0 kworker/2:0H
   29      0 watchdog/3
   30      0 migration/3
   31      0 ksoftirqd/3
   32      0 kworker/3:0
   33      0 kworker/3:0H
   34      0 khelper
   35      0 kdevtmpfs
   36      0 netns
   37      0 writeback
   38      0 kintegrityd
   39      0 bioset
   41      0 kblockd
   42      0 ata_sff
   43      0 khubd
   44      0 md
   45      0 devfreq_wq
   46      0 kworker/0:1
   47      0 kworker/1:1
   48      0 kworker/2:1
   50      0 khungtaskd
   51      0 kswapd0
   52      0 ksmd
   53      0 khugepaged
   54      0 fsnotify_mark
   55      0 ecryptfs-kthrea
   56      0 crypto
   68      0 kthrotld
   70      0 scsi_eh_0
   71      0 scsi_eh_1
   92      0 deferwq
   93      0 charger_manager
   94      0 kworker/1:2
   95      0 kworker/3:2
  149      0 kpsmoused
  155      0 jbd2/sda1-8
  156      0 ext4-rsv-conver
  316      0 jbd2/sda3-8
  317      0 ext4-rsv-conver
  565      0 kmemstick
  770      0 cfg80211
  818      0 hd-audio0
  853      0 kworker/2:2
  953      0 rpciod
  PID    VSZ
 1714      0 kauditd
11335      0 kworker/0:2
12202      0 kworker/u8:2
20228      0 kworker/u8:0
25529      0 kworker/u9:1
28305      0 kworker/u9:2
29822      0 lockd
 4919   4368 acpid
 4074   7136 ps
 6681  10232 dhclient
 4880  14540 getty
 4883  14540 getty
 4890  14540 getty
 4891  14540 getty
 4894  14540 getty
 6160  14540 getty
14486  15260 upstart-socket-
14489  15276 upstart-file-br
12009  18308 bash
12114  18308 bash
12289  18308 bash
 4075  19008 sort
11290  19140 atd
14483  19476 upstart-udev-br
11385  19972 bash
11553  19972 bash
11890  19972 bash
29503  21544 rpc.statd
 2847  23384 htop
  850  23420 rpcbind
29588  23480 rpc.idmapd
 4947  23656 cron
29833  24048 rpc.mountd
 5077  25324 hostapd
11301  26912 openvpn
    1  37356 init
 1152  39508 dbus-daemon
14673  43452 systemd-logind
14450  51204 systemd-udevd
 4921  61364 sshd
12008  67376 su
12113  67376 su
12288  67376 su
12007  67796 sudo
12112  67796 sudo
12287  67796 sudo
11336 107692 sshd
11384 107692 sshd
11502 107692 sshd
11841 107692 sshd
11552 108008 sshd
11889 108008 sshd
 1212 256228 rsyslogd
 1791 279832 polkitd
 4064 335684 whoopsie

Here is a full list of all running processes:

root@XanBox:~# ps aux
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.0  0.0  33644  1988 ?        Ss   Jul21   2:29 /sbin/init
root         2  0.0  0.0      0     0 ?        S    Jul21   0:03 [kthreadd]
root         3  0.0  0.0      0     0 ?        S    Jul21   1:04 [ksoftirqd/0]
root         5  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kworker/0:0H]
root         7  0.0  0.0      0     0 ?        S    Jul21  16:03 [rcu_sched]
root         8  0.0  0.0      0     0 ?        S    Jul21   4:08 [rcuos/0]
root         9  0.0  0.0      0     0 ?        S    Jul21   4:10 [rcuos/1]
root        10  0.0  0.0      0     0 ?        S    Jul21   4:30 [rcuos/2]
root        11  0.0  0.0      0     0 ?        S    Jul21   4:28 [rcuos/3]
root        12  0.0  0.0      0     0 ?        S    Jul21   0:00 [rcu_bh]
root        13  0.0  0.0      0     0 ?        S    Jul21   0:00 [rcuob/0]
root        14  0.0  0.0      0     0 ?        S    Jul21   0:00 [rcuob/1]
root        15  0.0  0.0      0     0 ?        S    Jul21   0:00 [rcuob/2]
root        16  0.0  0.0      0     0 ?        S    Jul21   0:00 [rcuob/3]
root        17  0.0  0.0      0     0 ?        S    Jul21   0:13 [migration/0]
root        18  0.0  0.0      0     0 ?        S    Jul21   0:08 [watchdog/0]
root        19  0.0  0.0      0     0 ?        S    Jul21   0:07 [watchdog/1]
root        20  0.0  0.0      0     0 ?        S    Jul21   0:13 [migration/1]
root        21  0.0  0.0      0     0 ?        S    Jul21   1:03 [ksoftirqd/1]
root        23  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kworker/1:0H]
root        24  0.0  0.0      0     0 ?        S    Jul21   0:07 [watchdog/2]
root        25  0.0  0.0      0     0 ?        S    Jul21   0:23 [migration/2]
root        26  0.0  0.0      0     0 ?        S    Jul21   1:01 [ksoftirqd/2]
root        28  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kworker/2:0H]
root        29  0.0  0.0      0     0 ?        S    Jul21   0:07 [watchdog/3]
root        30  0.0  0.0      0     0 ?        S    Jul21   0:23 [migration/3]
root        31  0.0  0.0      0     0 ?        S    Jul21   1:03 [ksoftirqd/3]
root        32  0.0  0.0      0     0 ?        S    Jul21   0:00 [kworker/3:0]
root        33  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kworker/3:0H]
root        34  0.0  0.0      0     0 ?        S<   Jul21   0:00 [khelper]
root        35  0.0  0.0      0     0 ?        S    Jul21   0:00 [kdevtmpfs]
root        36  0.0  0.0      0     0 ?        S<   Jul21   0:00 [netns]
root        37  0.0  0.0      0     0 ?        S<   Jul21   0:00 [writeback]
root        38  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kintegrityd]
root        39  0.0  0.0      0     0 ?        S<   Jul21   0:00 [bioset]
root        41  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kblockd]
root        42  0.0  0.0      0     0 ?        S<   Jul21   0:00 [ata_sff]
root        43  0.0  0.0      0     0 ?        S    Jul21   0:00 [khubd]
root        44  0.0  0.0      0     0 ?        S<   Jul21   0:00 [md]
root        45  0.0  0.0      0     0 ?        S<   Jul21   0:00 [devfreq_wq]
root        46  0.0  0.0      0     0 ?        S    Jul21  18:51 [kworker/0:1]
root        47  0.0  0.0      0     0 ?        S    Jul21   0:00 [kworker/1:1]
root        48  0.0  0.0      0     0 ?        S    Jul21   1:14 [kworker/2:1]
root        50  0.0  0.0      0     0 ?        S    Jul21   0:01 [khungtaskd]
root        51  0.4  0.0      0     0 ?        S    Jul21  95:51 [kswapd0]
root        52  0.0  0.0      0     0 ?        SN   Jul21   0:00 [ksmd]
root        53  0.0  0.0      0     0 ?        SN   Jul21   0:28 [khugepaged]
root        54  0.0  0.0      0     0 ?        S    Jul21   0:00 [fsnotify_mark]
root        55  0.0  0.0      0     0 ?        S    Jul21   0:00 [ecryptfs-kthrea]
root        56  0.0  0.0      0     0 ?        S<   Jul21   0:00 [crypto]
root        68  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kthrotld]
root        70  0.0  0.0      0     0 ?        S    Jul21   0:00 [scsi_eh_0]
root        71  0.0  0.0      0     0 ?        S    Jul21   0:00 [scsi_eh_1]
root        92  0.0  0.0      0     0 ?        S<   Jul21   0:00 [deferwq]
root        93  0.0  0.0      0     0 ?        S<   Jul21   0:00 [charger_manager]
root        94  0.0  0.0      0     0 ?        S    Jul21   1:05 [kworker/1:2]
root        95  0.0  0.0      0     0 ?        S    Jul21   1:08 [kworker/3:2]
root       149  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kpsmoused]
root       155  0.0  0.0      0     0 ?        S    Jul21   3:39 [jbd2/sda1-8]
root       156  0.0  0.0      0     0 ?        S<   Jul21   0:00 [ext4-rsv-conver]
root       316  0.0  0.0      0     0 ?        S    Jul21   1:28 [jbd2/sda3-8]
root       317  0.0  0.0      0     0 ?        S<   Jul21   0:00 [ext4-rsv-conver]
root       336  0.0  0.0  19476   512 ?        S    Jul21   0:07 upstart-udev-bridge --daemon
root       342  0.0  0.0  51228   468 ?        Ss   Jul21   0:00 /lib/systemd/systemd-udevd --daemon
root       565  0.0  0.0      0     0 ?        S<   Jul21   0:00 [kmemstick]
root       745  0.0  0.0  15364   252 ?        S    Jul21   0:06 upstart-socket-bridge --daemon
root       770  0.0  0.0      0     0 ?        S<   Jul21   0:00 [cfg80211]
root       818  0.0  0.0      0     0 ?        S<   Jul21   0:00 [hd-audio0]
root       850  0.0  0.0  23420    80 ?        Ss   Jul21   0:11 rpcbind
root       853  0.0  0.0      0     0 ?        S    Jul21   0:00 [kworker/2:2]
statd      872  0.0  0.0  21544     8 ?        Ss   Jul21   0:00 rpc.statd -L
root       953  0.0  0.0      0     0 ?        S<   Jul21   0:00 [rpciod]
root      1097  0.0  0.0  15276   364 ?        S    Jul21   0:06 upstart-file-bridge --daemon
message+  1152  0.0  0.0  39508  1448 ?        Ss   Jul21   1:40 dbus-daemon --system --fork
root      1157  0.0  0.0  23480     0 ?        Ss   Jul21   0:00 rpc.idmapd
root      1186  0.0  0.0  43736   984 ?        Ss   Jul21   1:13 /lib/systemd/systemd-logind
syslog    1212  0.0  0.0 256228   688 ?        Ssl  Jul21   1:41 rsyslogd
root      1714  0.0  0.0      0     0 ?        S    Jul21   0:00 [kauditd]
root      1791  0.0  0.0 279832  1312 ?        Sl   Jul21   1:16 /usr/lib/policykit-1/polkitd --no-debug
root      4880  0.0  0.0  14540     4 tty4     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty4
root      4883  0.0  0.0  14540     4 tty5     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty5
root      4890  0.0  0.0  14540     4 tty2     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty2
root      4891  0.0  0.0  14540     4 tty3     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty3
root      4894  0.0  0.0  14540     4 tty6     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty6
root      4919  0.0  0.0   4368     4 ?        Ss   Jul21   0:00 acpid -c /etc/acpi/events -s /var/run/acpid.socket
root      4921  0.0  0.0  61364   364 ?        Ss   Jul21   0:00 /usr/sbin/sshd -D
root      4947  0.0  0.0  23656   168 ?        Ss   Jul21   0:14 cron
root      5077  0.0  0.0  25324   648 ?        Ss   Jul21   0:34 /usr/sbin/hostapd -B -P /var/run/hostapd.pid /etc/hostapd/hostapd.conf
root      5192  0.0  0.0      0     0 ?        S    Jul21   0:00 [lockd]
root      5224  0.0  0.0  24048     4 ?        Ss   Jul21   0:00 /usr/sbin/rpc.mountd --manage-gids
root      6160  0.0  0.0  14540     4 tty1     Ss+  Jul21   0:00 /sbin/getty -8 38400 tty1
root      6681  0.0  0.0  10232     0 ?        Ss   11:07   0:00 dhclient -1 -v -pf /run/dhclient.eth0.pid -lf /var/lib/dhcp/dhclient.eth0.leases eth0
root      9452  0.0  0.0      0     0 ?        S    11:28   0:00 [kworker/u8:1]
root      9943  0.0  0.0      0     0 ?        S    11:42   0:00 [kworker/u8:0]
daemon   11290  0.0  0.0  19140   164 ?        Ss   11:59   0:00 atd
root     11301  0.2  0.1  26772  3436 ?        Ss   12:00   0:01 /usr/sbin/openvpn --writepid /var/run/openvpn.zanview.com.pid --status /var/run/openvpn.zanview.com.status 10 --cd /etc/openvpn --config /etc/openvpn/zanvie
root     11335  0.0  0.0      0     0 ?        S    12:01   0:00 [kworker/0:2]
root     11336  0.0  0.2 107692  4248 ?        Ss   12:01   0:00 sshd: deployer [priv]
deployer 11384  0.0  0.0 107692  1948 ?        S    12:01   0:00 sshd: deployer@pts/0
deployer 11385  0.0  0.1  19972  2392 pts/0    Ss+  12:01   0:00 -bash
root     11502  0.0  0.2 107692  4252 ?        Ss   12:01   0:00 sshd: deployer [priv]
deployer 11552  0.0  0.0 107692  1952 ?        S    12:01   0:00 sshd: deployer@pts/2
deployer 11553  0.0  0.1  19972  2388 pts/2    Ss   12:01   0:00 -bash
root     11841  0.0  0.2 107692  4248 ?        Ss   12:02   0:00 sshd: deployer [priv]
deployer 11889  0.0  0.1 108008  2280 ?        S    12:02   0:00 sshd: deployer@pts/3
deployer 11890  0.0  0.1  19972  2388 pts/3    Ss   12:02   0:00 -bash
root     12007  0.0  0.1  67796  2136 pts/3    S    12:02   0:00 sudo su -
root     12008  0.0  0.0  67376  2016 pts/3    S    12:02   0:00 su -
root     12009  0.0  0.1  18308  2228 pts/3    S+   12:02   0:00 -su
root     12112  0.0  0.1  67796  2136 pts/2    S    12:08   0:00 sudo su -
root     12113  0.0  0.0  67376  2012 pts/2    S    12:08   0:00 su -
root     12114  0.0  0.1  18308  2192 pts/2    S    12:08   0:00 -su
root     12180  0.0  0.0  15568  1160 pts/2    R+   12:09   0:00 ps aux
root     25529  0.0  0.0      0     0 ?        S<   Jul28   0:09 [kworker/u9:1]
root     28305  0.0  0.0      0     0 ?        S<   Aug05   0:00 [kworker/u9:2]

PS Mem Output

I also tried the ps_mem.py from https://github.com/pixelb/ps_mem

root@XanBox:~/ps_mem# python ps_mem.py
 Private  +   Shared  =  RAM used   Program

144.0 KiB +   9.5 KiB = 153.5 KiB   acpid
172.0 KiB +  29.5 KiB = 201.5 KiB   atd
248.0 KiB +  35.0 KiB = 283.0 KiB   cron
272.0 KiB +  84.0 KiB = 356.0 KiB   upstart-file-bridge
276.0 KiB +  84.5 KiB = 360.5 KiB   upstart-socket-bridge
280.0 KiB + 102.5 KiB = 382.5 KiB   upstart-udev-bridge
332.0 KiB +  54.5 KiB = 386.5 KiB   rpc.idmapd
368.0 KiB +  91.5 KiB = 459.5 KiB   rpcbind
388.0 KiB + 251.5 KiB = 639.5 KiB   systemd-logind
668.0 KiB +  43.5 KiB = 711.5 KiB   hostapd
576.0 KiB + 157.5 KiB = 733.5 KiB   systemd-udevd
676.0 KiB +  65.5 KiB = 741.5 KiB   rpc.mountd
604.0 KiB + 163.0 KiB = 767.0 KiB   rpc.statd
908.0 KiB +  62.5 KiB = 970.5 KiB   dbus-daemon [updated]
932.0 KiB + 117.0 KiB =   1.0 MiB   getty [updated] (6)
  1.0 MiB +  69.5 KiB =   1.1 MiB   openvpn
  1.0 MiB + 137.0 KiB =   1.2 MiB   polkitd
  1.5 MiB + 202.0 KiB =   1.7 MiB   htop
  1.4 MiB + 306.5 KiB =   1.7 MiB   whoopsie
  1.4 MiB + 279.0 KiB =   1.7 MiB   su (3)
  1.5 MiB + 268.5 KiB =   1.8 MiB   sudo (3)
  2.2 MiB +  11.5 KiB =   2.3 MiB   dhclient
  3.9 MiB + 741.0 KiB =   4.6 MiB   bash (6)
  5.3 MiB + 254.5 KiB =   5.5 MiB   init
  2.7 MiB +   3.3 MiB =   6.1 MiB   sshd (7)
 18.1 MiB +  56.5 KiB =  18.2 MiB   rsyslogd
---------------------------------
                         53.7 MiB
=================================

Slabtop Output

I also tried slabtop:

root@XanBox:~# slabtop -sc
 Active / Total Objects (% used)    : 131306 / 137558 (95.5%)
 Active / Total Slabs (% used)      : 3888 / 3888 (100.0%)
 Active / Total Caches (% used)     : 63 / 105 (60.0%)
 Active / Total Size (% used)       : 27419.31K / 29580.53K (92.7%)
 Minimum / Average / Maximum Object : 0.01K / 0.21K / 8.00K

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
  8288   7975  96%    0.57K    296       28      4736K inode_cache
 14259  12858  90%    0.19K    679       21      2716K dentry
  2384   1943  81%    0.96K    149       16      2384K ext4_inode_cache
 20916  20494  97%    0.11K    581       36      2324K sysfs_dir_cache
   624    554  88%    2.00K     39       16      1248K kmalloc-2048
   195    176  90%    5.98K     39        5      1248K task_struct
  6447   6387  99%    0.19K    307       21      1228K kmalloc-192
  2128   1207  56%    0.55K     76       28      1216K radix_tree_node
   768    761  99%    1.00K     48       16       768K kmalloc-1024
   176    155  88%    4.00K     22        8       704K kmalloc-4096
  1100   1100 100%    0.63K     44       25       704K proc_inode_cache
  1008   1008 100%    0.66K     42       24       672K shmem_inode_cache
  2640   2262  85%    0.25K    165       16       660K kmalloc-256
   300    300 100%    2.06K     20       15       640K sighand_cache
  5967   5967 100%    0.10K    153       39       612K buffer_head
  1152   1053  91%    0.50K     72       16       576K kmalloc-512
  3810   3810 100%    0.13K    127       30       508K ext4_allocation_context
    60     60 100%    8.00K     15        4       480K kmalloc-8192
   225    225 100%    2.06K     15       15       480K idr_layer_cache
  7616   7324  96%    0.06K    119       64       476K kmalloc-64
   700    700 100%    0.62K     28       25       448K sock_inode_cache
   252    252 100%    1.75K     14       18       448K TCP
  8925   8544  95%    0.05K    105       85       420K shared_policy_node
  3072   2351  76%    0.12K     96       32       384K kmalloc-128
   360    360 100%    1.06K     12       30       384K signal_cache
   432    337  78%    0.88K     24       18       384K mm_struct

Other

I also tried scanning for a rootkit with rkhunter - it found nothing. And I tried to sync and dump cache with:

sync; sync; sync; echo 3 > /proc/sys/vm/drop_caches

It made no difference also.

I also tried to force swap or disable swap with:

sudo sysctl -w vm.swappiness=100
sudo swapoff /dev/sda2

I also tried using htop and sorting by memory and it is not showing where the memory is going either. The kernel version is Linux 3.13.0-40-generic #69-Ubuntu SMP.

Dmesg output: http://pastie.org/9558255 smem output: http://pastie.org/9558290

Conclusion

What is going on? - Where is all the memory going? - How do I find out?

10

4 Answers 4

8
+50

My conclusion is it is a kernel memory leak somewhere in the Linux kernel, this is why none of the userspace tools are able to show where memory is being leaked. Maybe it is related to this question: https://serverfault.com/questions/670423/linux-memory-usage-higher-than-sum-of-processes

I upgraded the kernel version from 3.13 to 3.19 and it seems the memory leak has stopped! - I will report back if I see a leak again.

It would still be useful to have some easy/easier way to see how much memory is used for different parts of the Linux kernel. It is still a mystery what was causing the leak in 3.13.

2
4

Story

I can reproduce your issue using ZFS on Linux.

Here is a server called node51 with 20GB of RAM. I marked 16GiB of RAM to be allocatable to the ZFS adaptive replacement cache (ARC):

root@node51 [~]# echo 17179869184 > /sys/module/zfs/parameters/zfs_arc_max
root@node51 [~]# grep c_max /proc/spl/kstat/zfs/arcstats
c_max                           4    17179869184

Then, I read a 45GiB file using Pipe Viewer in my ZFS pool zeltik to fill up the ARC:

root@node51 [~]# pv /zeltik/backup-backups/2014.04.11.squashfs > /dev/zero
  45GB 0:01:20 [ 575MB/s] [==================================>] 100%

Now look at the free memory:

root@node51 [~]# free -m
             total       used       free     shared    buffers     cached
Mem:         20013      19810        203          1         51         69
-/+ buffers/cache:      19688        324
Swap:         7557          0       7556

Look!

51MiB in buffers
69MiB in cache
120MiB in both
19688MiB of RAM in use, including buffers and cache
19568MiB of RAM in use, excluding buffers and cache

The Python script that you referenced reports that applications are only using a small amount of RAM:

root@node51 [~]# python ps_mem.py
 Private  +   Shared  =  RAM used       Program

148.0 KiB +  54.0 KiB = 202.0 KiB       acpid
176.0 KiB +  47.0 KiB = 223.0 KiB       swapspace
184.0 KiB +  51.0 KiB = 235.0 KiB       atd
220.0 KiB +  57.0 KiB = 277.0 KiB       rpc.idmapd
304.0 KiB +  62.0 KiB = 366.0 KiB       irqbalance
312.0 KiB +  64.0 KiB = 376.0 KiB       sftp-server
308.0 KiB +  89.0 KiB = 397.0 KiB       rpcbind
300.0 KiB + 104.5 KiB = 404.5 KiB       cron
368.0 KiB +  99.0 KiB = 467.0 KiB       upstart-socket-bridge
560.0 KiB + 180.0 KiB = 740.0 KiB       systemd-logind
724.0 KiB +  93.0 KiB = 817.0 KiB       dbus-daemon
720.0 KiB + 136.0 KiB = 856.0 KiB       systemd-udevd
912.0 KiB + 118.5 KiB =   1.0 MiB       upstart-udev-bridge
920.0 KiB + 180.0 KiB =   1.1 MiB       rpc.statd (2)
  1.0 MiB + 129.5 KiB =   1.1 MiB       screen
  1.1 MiB +  84.5 KiB =   1.2 MiB       upstart-file-bridge
960.0 KiB + 452.0 KiB =   1.4 MiB       getty (6)
  1.6 MiB + 143.0 KiB =   1.7 MiB       init
  5.1 MiB +   1.5 MiB =   6.5 MiB       bash (3)
  5.7 MiB +   5.2 MiB =  10.9 MiB       sshd (8)
 11.7 MiB + 322.0 KiB =  12.0 MiB       glusterd
 27.3 MiB +  99.0 KiB =  27.4 MiB       rsyslogd
 67.4 MiB + 453.0 KiB =  67.8 MiB       glusterfsd (2)
---------------------------------
                        137.4 MiB
=================================

19568MiB - 137.4MiB ≈ 19431MiB of unaccounted RAM

Explanation

The 120MiB of buffers and cache used that you saw in the story above account for the kernel's efficient behavior of caching data sent to or received from an external device.

The first row, labeled Mem, displays physical memory utilization, including the amount of memory allocated to buffers and caches. A buffer, also called buffer memory, is usually defined as a portion of memory that is set aside as a temporary holding place for data that is being sent to or received from an external device, such as a HDD, keyboard, printer or network.

The second line of data, which begins with -/+ buffers/cache, shows the amount of physical memory currently devoted to system buffer cache. This is particularly meaningful with regard to application programs, as all data accessed from files on the system that are performed through the use of read() and write() system calls pass through this cache. This cache can greatly speed up access to data by reducing or eliminating the need to read from or write to the HDD or other disk.

Source: http://www.linfo.org/free.html

Now how do we account for the missing 19431MiB?

In the free -m output above, the 19688MiB "used" in "-/+ buffers/cache" comes from this formula:

(kb_main_used) - (buffers_plus_cached) =
(kb_main_total - kb_main_free) - (kb_main_buffers + kb_main_cached)

kb_main_total:   MemTotal from /proc/meminfo
kb_main_free:    MemFree  from /proc/meminfo
kb_main_buffers: Buffers  from /proc/meminfo
kb_main_cached:  Cached   from /proc/meminfo

Source: procps/free.c and procps/proc/sysinfo.c

(If you do the numbers based on my free -m output, you'll notice that 2MiB aren't accounted for, but that's because of rounding errors introduced by this code: #define S(X) ( ((unsigned long long)(X) << 10) >> shift))

The numbers don't add up in /proc/meminfo, either (I didn't record /proc/meminfo when I ran free -m, but we can see from your question that /proc/meminfo doesn't show where the missing RAM is), so we can conclude from the above that /proc/meminfo doesn't tell the whole story.

In my testing conditions, I know as a control that ZFS on Linux is responsible for the high RAM usage. I told its ARC that it could use up to 16GiB of the server's RAM.

ZFS on Linux isn't a process. It's a kernel module.

From what I've found so far, the RAM usage of a kernel module wouldn't show up using process information tools because the module isn't a process.

Troubleshooting

Unfortunately, I don't know enough about Linux to offer you a way to build a list of how much RAM non-process components (like the kernel and its modules) are using.

At this point, we can speculate, guess, and check.

You provided a dmesg output. Well-designed kernel modules would log some of their details to dmesg.

After looking through dmesg, one item stood out to me: FS-Cache

FS-Cache is part of the cachefiles kernel module and relates to the package cachefilesd on Debian and Red Hat Enterprise Linux.

Perhaps some time ago, you configured FS-Cache on a RAM disk to reduce the impact of network I/O as your server analyzes the video data.

Try disabling any suspicious kernel modules that could be eating up RAM. They can probably be disabled with blacklist in /etc/modprobe.d/, followed by a sudo update-initramfs -u (commands and locations may vary by Linux distribution).

Conclusion

A memory leak is eating up 8MB/hr of your RAM and won't release the RAM, seemingly no matter what you do. I was not able to determine the source of your memory leak based on the information that you provided, nor was I able to offer a way to find that memory leak.

Someone who is more experienced with Linux than I will need to provide input on how we can determine where the "other" RAM usage is going.

I have started a bounty on this question to see if we can get a better answer than "speculate, guess, and check".

1
  • Thank you for that and thank you for the bounty! Just a note, I did not enable any caching, the video stream is analysed live and there is no network storage at all (no NFS, no SMB, FS_Cache is not used and I do not have cachefilesd installed). I tried looking at all loaded kernel modules and unloading anything unused already which also did not help :( - I suspect maybe the leak is in the networking driver or something - but I guess both mine and your question is how to troubleshoot this outside of guessing.
    – Hackeron
    Commented Aug 23, 2015 at 14:19
1

Do you change the Swapiness of your Kernel manualy or disable it?

you can whatch you current swappyness-level with

cat /proc/sys/vm/swappiness

You could try to force your kernel to swap aggressively with

sudo sysctl -w vm.swappiness=100

if this decrease you problems find a good value between 1 and 100, fitting your requirement.

1
  • swappiness has nothing to do with it - even if I turn swap off completely, I still have a situation where I've killed every user process and still all the memory is used up.
    – Hackeron
    Commented Sep 16, 2014 at 8:35
-3

You are not quite right – yes your free –m command is showing free 220MB but it is also showing that 1771MB is used as buffers. Buffers and Cached is memory used by the kernel to optimize access to slow access data, usually disks. So you should consider all memory marked as buffers as free memory because kernel can take it back whenever it is required.

See: https://serverfault.com/questions/23433/in-linux-what-is-the-difference-between-buffers-and-cache-reported-by-the-f

1
  • I updated the answer to take this into account. I am sorry to say you are fundamentally misunderstanding the output from the "free" command.
    – Hackeron
    Commented Aug 23, 2015 at 11:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .