networking - 使用 MRTG 进行带宽监控，没有来自交换机堆栈的数据

Question

我正在尝试建立一个监控系统，并被要求在其中添加我们所有的上行链路。我们的大多数交换机由大约 7 个交换机组成。这些堆栈不返回任何数据。

我在单个交换机上运行 MRTG 并返回漂亮的图表，对于我们的 SAN 交换机也是如此。但大堆栈只返回空日志和空图。它似乎确实看到了哪些端口正在使用，因为它不会为未使用的端口返回它。

有人知道出了什么问题吗？它在具有最新 MRTG 软件包的 ubuntu 服务器上运行。

Answer 1

There are several reasons why you might not be getting any data in the MRTG graphs. In order to find out more, you should run MRTG interactively from the command line, and look at the messages produced - is it telling you any error messages? Can you use snmpwalk to successfully query the stack from the commandline?

Possible reasons -

1. You have the incorrect SNMP community string to access the devices
2. You have no network access to these devices (routing, firewall, ACL on devices)
3. You have the wrong SNMP version (use SNMPv2 where possible)
4. The ports are GigE and your device will only report in SNMPv2 -- and you're using SNMPv1
5. The stacks are so large that the SNMP times out before returning information. Increase the timeout.
6. You have added cards or items to the stacks, and this has reenumerated the ports but you have not regenerated the MRTG cfg file, so MRTG is querying nonexistent items. Re-run cfgmaker.
7. The MaxBytes setting in the cfg file is too low, and the returned data are being ignored. This can happen with some devices that return invalid port details for GigE ports and cause cfgmaker to generate incorrect configurations.

You can set the SNMP version and timeout options in the Target[] definition; see the documentation in http://oss.oetiker.ch/mrtg/doc/mrtg-reference.en.html