MRTG和/或rrdtool - 一分钟间隔不工作

Question

我在Ubuntu 16.04上测试MRTG 2.17.4。我将其配置为使用rrdtool来记录数据。我的/etc/mrtg.cfg中有“Interval：1”。我编辑在/etc/cron.d/mrtg cron任务是：

*/1 * * * * root if [ -x /usr/bin/mrtg ] && [ -r /etc/mrtg.cfg ] && [ -d "$(grep '^[[:space:]]*[^#]*[[:space:]]*WorkDir' /etc/mrtg.cfg | awk '{ print $NF }')" ]; then mkdir -p /var/log/mrtg ; env LANG=C /usr/bin/mrtg /etc/mrtg.cfg --debug="time,log" 2>&1 | tee -a /var/log/mrtg/mrtg.log ; fi 

请注意，我添加调试选项“时，登录”，以帮助排除故障。在我的/var/log/mrtg/mrtg.log中，我发现数据按预期每隔一分钟记录一次，但不会每分钟计算一次吞吐量。相反，它每5分钟做一次。对于其他时间，它只是说“有???/???”

--time: prog start Sat Nov 12 21:54:01 2016 
--time: loop start Sat Nov 12 21:54:01 2016 
--time: snmp read start Sat Nov 12 21:54:01 2016 
--time: target loop start Sat Nov 12 21:54:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479012841:171337092:172912456') 
--log: got: ???/??? 
--time: prog start Sat Nov 12 21:55:01 2016 
--time: loop start Sat Nov 12 21:55:01 2016 
--time: snmp read start Sat Nov 12 21:55:01 2016 
--time: target loop start Sat Nov 12 21:55:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479012901:171686542:173262414') 
--log: got: 5804.75277777778/5813.21944444444 
--time: prog start Sat Nov 12 21:56:02 2016 
--time: loop start Sat Nov 12 21:56:02 2016 
--time: snmp read start Sat Nov 12 21:56:02 2016 
--time: target loop start Sat Nov 12 21:56:02 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479012962:172034542:173610922') 
--log: got: ???/??? 
--time: prog start Sat Nov 12 21:57:01 2016 
--time: loop start Sat Nov 12 21:57:01 2016 
--time: snmp read start Sat Nov 12 21:57:01 2016 
--time: target loop start Sat Nov 12 21:57:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479013021:172376742:173953630') 
--log: got: ???/??? 
--time: prog start Sat Nov 12 21:58:01 2016 
--time: loop start Sat Nov 12 21:58:01 2016 
--time: snmp read start Sat Nov 12 21:58:01 2016 
--time: target loop start Sat Nov 12 21:58:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479013081:172724742:174302138') 
--log: got: ???/??? 
--time: prog start Sat Nov 12 21:59:01 2016 
--time: loop start Sat Nov 12 21:59:01 2016 
--time: snmp read start Sat Nov 12 21:59:01 2016 
--time: target loop start Sat Nov 12 21:59:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479013141:173077092:174654996') 
--log: got: ???/??? 
--time: prog start Sat Nov 12 22:00:02 2016 
--time: loop start Sat Nov 12 22:00:02 2016 
--time: snmp read start Sat Nov 12 22:00:02 2016 
--time: target loop start Sat Nov 12 22:00:02 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479013202:173426542:175004954') 
--log: got: 5781.22263205829/5789.66200182149 
--time: prog start Sat Nov 12 22:01:01 2016 
--time: loop start Sat Nov 12 22:01:01 2016 
--time: snmp read start Sat Nov 12 22:01:01 2016 
--time: target loop start Sat Nov 12 22:01:01 2016 
--log: RRDs::tune(/var/mrtg/192.168.1.40_24.rrd -a ds0:125000000 -a ds1:125000000 -d ds0:COUNTER -d ds1:COUNTER) 
--log: RRDs::update(/var/mrtg/192.168.1.40_24.rrd, '1479013261:173768742:175347662') 
--log: got: ???/??? 

我并不特别担心它不会每分钟计算一次速度，但是我需要将数据存储在rrd数据库中。但是当我尝试使用rrdtool获取时，它仍然显示5分钟的数据..我在哪里做错了？

rrdtool fetch 192.168.1.40_24.rrd AVERAGE -a -s -1h -r 1m -a 
          ds0     ds1 

1479010500: 1.0150000000e+04 1.0158466667e+04 
1479010800: 1.0150000000e+04 1.0158466667e+04 
1479011100: 1.0145166667e+04 1.0153661556e+04 
1479011400: 1.0021447006e+04 1.0029884972e+04 
1479011700: 5.8127796610e+03 5.8212468060e+03 
1479012000: 5.7859754098e+03 5.7944147796e+03 
1479: 5.8089454941e+03 5.8174389793e+03 
1479012600: 5.8006990960e+03 5.8091662411e+03 
1479012900: 5.8047527778e+03 5.8132194444e+03 
1479013200: 5.7812226321e+03 5.7896620018e+03 
1479013500: 5.7946912568e+03 5.8031847421e+03 
1479013800: 5.8049794444e+03 5.8134465895e+03 
1479014100: -nan -nan 

谢谢！ Difan

Answer 1

它可能有助于查看RRT文件中的MRTG配置文件和rrdtool info的输出。

但是，可能的原因是您最初以5分钟的时间间隔创建RRD文件，然后稍后更改为1分钟的时间间隔。如果你这样做，那么MRTG将不会重新创建RRD文件，尽管它会改变为每1分钟收集一次数据。 Mrtg只能在创建后修改RRD的MaxBytes;其他设置（RRDRowCount，Interval）只能在初始创建时使用，并且随后在MRTG cfg文件中更改它们不会导致RRD文件更改。

如果您在RRD文件上使用rrdtool info，则可能会看到它具有5分钟（300s）的步长。改变这种情况的唯一方法是删除RRD文件（丢失历史数据）并允许MRTG以cfg文件中指定的60秒间隔重新创建它。请注意，如果您必须保留5分钟RRD文件中的许多有价值的历史数据，则可以使用第三方实用程序将其转储为XML并重新创建新的RRD，但该任务是有点复杂。寻找rrdmigrate或rrdmerge。最新的RRDtool在这方面也有一些能力。