2

I have got a Ubuntu server guest virtual machine (installed via virt-install) which runs Nagios Core monitoring but there is a major issue with the monitoring (i.e. ping packet loss / RTA or connection time outs indicated intermittently in the Nagios web interface).

I am monitoring 19 hosts (UP/DOWN) and ping service only with default intervals/rechecks etc.

The Ubuntu server VM has 1 CPU (default settings), 5GB storage and 1 GB memory (there is plenty resource free) but uses all the default virt-install network settings*.

There are a number of other (Debian) guest VMs on the VM host.

ping (from command line) is slightly slower on this VM and sometimes hangs for several seconds (this doesn't happen on other VMs).

*EDIT: I thought it may be a network interface issue so changed it to use virtio (in VM XML interface configuration) but even after that the problem persists.

It seems to be a network (config) issue on this particular VM.

PING INVESTIGATION:

  1. From this Ubuntu guest VM with network issue:
--- X.X.X.X ping statistics ---
100 packets transmitted, 63 received, 37% packet loss, time 101341ms
rtt min/avg/max/mdev = 0.352/0.491/0.637/0.056 ms
  1. From another (Debian) guest VM:
--- X.X.X.X ping statistics ---
100 packets transmitted, 100 received, 0% packet loss, time 101371ms
rtt min/avg/max/mdev = 0.278/0.386/0.505/0.056 ms

Ping from the problem guest VM to the other guest VM:

--- X.X.X.X ping statistics ---
100 packets transmitted, 100 received, 0% packet loss, time 101354ms
rtt min/avg/max/mdev = 0.345/0.531/0.798/0.092 ms

There is no apparent IP conflict either.

2one
  • 131
  • 5
  • 1
    Are you sure whether that is a Nagios problem? Maybe there is a networking problem, and Nagios is correctly reporting that. – Hauke Laging Apr 23 '20 at 17:55
  • Yes i don't think it's nessicarily a Nagios problem. I'm wondering if it's the default virt-install network provision plus the fact there are other VM on the host... – 2one Apr 23 '20 at 18:04
  • 1
    Then do the check operation (e.g. send a ping) "manually" (often enough for a solid statement) and see what happens. – Hauke Laging Apr 23 '20 at 18:08
  • I've added some more detail. Thanks. – 2one Apr 24 '20 at 11:00
  • 1
    So not Nagios-related. You should make sure that there is no IP conflict. Take the problem VM offline and try to ping its IP afterwards (after flushing the ARP cache). In addition: What happens if the other VM pings the problem VM? – Hauke Laging Apr 24 '20 at 11:08

0 Answers0