memory usage detected wrong for the host system

[Lukas Schreiner]

Due to bug #379 i had to use older version or the master branch. Right now i'm using the master branch with ref. 27210ed610ba

To the bug. My config:

check system localhost.local
    if loadavg (1min) > 4 for 2 cycles then alert
    if loadavg (5min) > 2 for 2 cycles then alert
    if memory usage > 95% for 2 cycles then alert
    if cpu usage (user) > 90% for 4 cycles then alert
    if cpu usage (system) > 80% for 4 cycles then alert
    if cpu usage (wait) > 80% for 4 cycles then alert
    if swap usage > 85% for 1 cycles then alert

If the memory usage is > 95% i want an alert. But it is now always complaining:

System 'localhost.local'
  status                       Resource limit matched
  monitoring status            Monitored
  monitoring mode              active
  on reboot                    start
  load average                 [0.00] [0.07] [0.09]
  cpu                          2.4%us 1.8%sy 0.4%wa
  memory usage                 16.0 EB [1717986918400.0%]
  swap usage                   0 B [0.0%]
  uptime                       171d 16h 52m
  boot time                    Wed, 10 Feb 2016 03:17:47
  data collected               Sat, 30 Jul 2016 21:09:41

16.0 EB [1717986918400.0%] is a little bit high. Output of free:

             total       used       free     shared    buffers     cached
Mem:       1048576    1048576          0      97040          0     534608
-/+ buffers/cache:     513968     534608
Swap:      1048576          0    1048576

Note: this server is running via Linux-vServer - on my KVM servers its running without problems.

I had before the version 1.15 of monit. With that version, it worked out of the box.

If you need further information, let me know.

Thanks!

[Tildeslash]

Please try with master head and let us know if the problem exist.

[Lukas Schreiner]

It is already master head.

[Lukas Schreiner]

OK.. have to correct me. Issue is already in monit version 5.15. But the output is different:

The Monit daemon 5.15 uptime: 0m 

System 'localhost.local'
  status                            Running
  monitoring status                 Monitored
  load average                      [0.70] [0.34] [0.15]
  cpu                               0.0%us 0.0%sy 0.0%wa
  memory usage                      16.0 ZB [0.0%]
  swap usage                        0 B [0.0%]
  data collected                    Sat, 30 Jul 2016 23:20:14

Same for 5.14.

Monit 5.10 it was negative:

The Monit daemon 5.10 uptime: 0m 

System 'localhost.local'
  status                            Running
  monitoring status                 Monitored
  load average                      [0.68] [0.48] [0.27]
  cpu                               0.0%us 0.0%sy 0.0%wa
  memory usage                      16.0 ZB [-214748364.8%]
  swap usage                        0.0 B [0.0%]
  data collected                    Sat, 30 Jul 2016 23:28:20

[Lukas Schreiner]

# cat /proc/meminfo
MemTotal:        1048576 kB
MemFree:               0 kB
Buffers:               0 kB
Cached:           454544 kB
SwapCached:       435112 kB
Active:          8960756 kB
Inactive:        3675180 kB
Active(anon):    3814280 kB
Inactive(anon):  1209144 kB
Active(file):    5146476 kB
Inactive(file):  2466036 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       1048576 kB
SwapFree:        1048576 kB
Dirty:              1948 kB
Writeback:             0 kB
AnonPages:       4557128 kB
Mapped:           320948 kB
Shmem:             97008 kB
Slab:            2911940 kB
SReclaimable:    2498852 kB
SUnreclaim:       413088 kB
KernelStack:        9728 kB
PageTables:       100776 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:    25187216 kB
Committed_AS:    9964672 kB
VmallocTotal:   34359738367 kB
VmallocUsed:       51564 kB
VmallocChunk:   34359612116 kB
HardwareCorrupted:     0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:        2040 kB
DirectMap2M:     2095104 kB
DirectMap1G:    14680064 kB

[Tildeslash]

I checked in a small patch to clarify types. Please try master head again

[Lukas Schreiner]

Hi, unchanged:

root@www /u/l/s/monit# monit status localhost.local
Monit 5.19 uptime: 1m

System 'localhost.local'
  status                       Running
  monitoring status            Monitored
  monitoring mode              active
  on reboot                    start
  load average                 [0.41] [0.23] [0.09]
  cpu                          0.0%us 0.0%sy 0.0%wa
  memory usage                 16.0 EB [1717986918400.0%]
  swap usage                   0 B [0.0%]
  uptime                       172d 10h 41m
  boot time                    Wed, 10 Feb 2016 03:17:47
  data collected               Sun, 31 Jul 2016 14:59:38

[Tildeslash]

Please run Monit in debug mode monit -Iv and let me know if you can see this error message in the output, system statistic error -- cannot get real memory amount. If you see any other error messages, please let me know as well. Also try to run as root if you haven't already. I suspect that Linux-vServer either has a bug or some security type of protection, do you know? Is your server 64 bits BTW?

[Lukas Schreiner]

Thanks for the reply. As far as i see, there is no statistics error. Also no other error. Monit is running as root.

I understand what you mean regarding the Linux-vServer. I also had such an idea. At least we have the issue, that we cannot do things, which are near to the kernel (no own kernel, no access to driver options or to hardware, no tap/tun, no own network device; its all isolated). Of course, basic files are there (like this meminfo, cpuinfo,..)

regarding 64 bit:

# uname -m
x86_64

[Tildeslash]

Duplicate of #385.