Ad Widget

Collapse

Недополучение данных с нагруженного хос&

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • yanda.a
    Junior Member
    • Jan 2016
    • 11

    #1

    Недополучение данных с нагруженного хос&

    Добрый день!

    Проблема в следующем. Есть сервер, на котором установлен Zabbix Agent, используется активное получение данных. При высокой нагрузке на сервер теряются данные:
    Code:
    CPU | sys    1014%  | user    498%  | irq      33%  |               |               | idle    860%  | wait      0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      40%  | user     36%  | irq       0%  |               |               | idle     24%  | cpu017 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      45%  | user     30%  | irq       0%  |               |               | idle     25%  | cpu013 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      40%  | user     32%  | irq       0%  |               |               | idle     28%  | cpu022 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      44%  | user     27%  | irq       0%  |               |               | idle     29%  | cpu018 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      43%  | user     26%  | irq       0%  |               |               | idle     31%  | cpu019 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      43%  | user     26%  | irq       0%  |               |               | idle     31%  | cpu014 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      43%  | user     24%  | irq       0%  |               |               | idle     33%  | cpu021 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      39%  | user     27%  | irq       0%  |               |               | idle     33%  | cpu020 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      43%  | user     23%  | irq       0%  |               |               | idle     34%  | cpu023 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      40%  | user     25%  | irq       0%  |               |               | idle     35%  | cpu016 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      46%  | user     19%  | irq       0%  |               |               | idle     35%  | cpu009 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      44%  | user     20%  | irq       0%  |               |               | idle     35%  | cpu008 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      44%  | user     20%  | irq       0%  |               |               | idle     36%  | cpu015 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      45%  | user     19%  | irq       0%  |               |               | idle     36%  | cpu010 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      45%  | user     17%  | irq       0%  |               |               | idle     38%  | cpu011 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      37%  | user     24%  | irq       0%  |               |               | idle     39%  | cpu012 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      44%  | user     11%  | irq       5%  |               |               | idle     40%  | cpu002 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      40%  | user     16%  | irq       4%  |               |               | idle     40%  | cpu007 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      42%  | user     11%  | irq       5%  |               |               | idle     42%  | cpu000 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      40%  | user     13%  | irq       5%  |               |               | idle     42%  | cpu004 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      39%  | user     15%  | irq       4%  |               |               | idle     42%  | cpu005 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      41%  | user     12%  | irq       4%  |               |               | idle     43%  | cpu001 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      42%  | user     11%  | irq       4%  |               |               | idle     43%  | cpu003 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    cpu | sys      42%  | user     12%  | irq       2%  |               |               | idle     44%  | cpu006 w  0%  |               | steal     0%  |               | guest     0%  | curf 2.50GHz  | curscal  99%  |
    CPL | avg1   20.19  |               | avg5   13.27  | avg15   9.08  |               |               |               | csw   373478  |               | intr   85539  |               |               | numcpu    24  |
    При такой нагрузке мы видим потери как на скриншоте. Zabbix Agent запускается с nice'ом:
    Code:
    root@server:/var/log/atop # ps -ax -o nice,command | grep zabbix
    -16 /usr/local/sbin/zabbix_agentd
    -16 zabbix_agentd: collector [idle 1 sec] (zabbix_agentd)
    -16 zabbix_agentd: listener #1 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #2 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #3 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #4 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #5 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #6 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #7 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: listener #8 [waiting for connection] (zabbix_agentd)
    -16 zabbix_agentd: active checks #1 [processing active checks] (zabbix_agentd)
    -16 zabbix_agentd: active checks #2 [processing active checks] (zabbix_agentd)
    ОС:
    Code:
    root@server:/var/log/atop # uname -a
    FreeBSD server10.freeteam.org 9.3-RELEASE-p33 FreeBSD 9.3-RELEASE-p33 #0: Wed Jan 13 17:55:39 UTC 2016     [email protected]:/usr/obj/usr/src/sys/GENERIC  amd64
    Настройки Zabbix Agent:
    Code:
    root@server:/usr/local/etc/zabbix24 # grep -vE '^#|^$' zabbix_agentd.conf 
    LogFile=/var/log/zabbix/zabbix_agentd.log
    LogFileSize=0
    DebugLevel=4
    EnableRemoteCommands=1
    LogRemoteCommands=1
    Server=XXX.XXX.XXX.XXX
    StartAgents=8
    ServerActive=XXX.XXX.XXX.XXX
    HostnameItem=system.hostname
    HostMetadataItem=system.uname
    BufferSize=4096
    MaxLinesPerSecond=1000
    Timeout=30
    Include=/usr/local/etc/zabbix24/zabbix_agentd.conf.d/*_freebsd.conf
    Собственно, как быть в данной ситуации? Дело в том, что с munin не обнаружено подобных проблем, может есть возможность как-то оптимизировать работу Zabbix'а?
    Attached Files
    Last edited by yanda.a; 28-01-2016, 10:18. Reason: Add screenshot
Working...