Troubleshooting Help Requested
-
On a specific host I have several VMs, several of which are Windows based, others that are CentOS 7 based.
To provide some insight on the Windows VMs the network is fully functional, internet, can ping via dns or ip etc etc.
From the CentOS box, I no longer have the ability to ping via DNS or IP to the internet. This used to work without issue and nothing has been customized on the box.
- The box does have a valid IP address, gateway and DNS servers.
- I've rebooted
- I can't reboot the hypervisor (plus the other VMs are working just fine).
I found out that this box was having issues when I went to perform updates on this one CentOS box and trying to figure out what has occurred to cause this.
-
@DustinB3403 My first suspicion is always /etc/resolve.conf getting defaulted from something being stupid (Network Manager, I'm looking at you!) How did you verify that DNS is working? Can you ping the gateway? If it's installed, maybe review the settings with nmcli.
That's so annoying.
-
Everything looks good here.
-
A separate, yet working VM on a different host.
-
@DustinB3403 You can ping the gateway and dns servers, right? How about nslookup internet sites?
Guess I get to ask all the annoying, simple questions, this morning.
-
I can ping the gateway.
I can ping things on the network, by both IP and name. From both the working and non working VM.
-
Have you tried just disconnecting (wait 10 sec) and then reconnecting the NIC in the Hypervisor?
Had to do this once or twice in the past (Ubuntu). No idea but it worked.The only other thing was to verify that the MAC of NIC in Hypervisor matched MAC in VM. That has switched on me in the past too.
-
@travisdh1 said in Troubleshooting Help Requested:
@DustinB3403 You can ping the gateway and dns servers, right? How about nslookup internet sites?
Guess I get to ask all the annoying, simple questions, this morning.
nslookup www.google.com bash: nslookup: command not foundNot installed on the machine, and I can't install it because I "apparently don't have internet" yet on the working machine.
I can, find bind-utils, install bind-utils and nslookup google.com
nslookup www.google.com Server: 192.168.2.16 Address: 192.168.2.16#53 Non-authoritative answer: Name: www.google.com Address: 172.217.10.228 -
On the Hypervisor with the "not connected vm" all VMs on the host use the same virtual interface for internet access.
So there isn't a hardware issue there, not that I can think of anyways that would affect only the 1 guest.
-
So if you traceroute to 8.8.8.8, what do you get?
-
@scottalanmiller said in Troubleshooting Help Requested:
So if you traceroute to 8.8.8.8, what do you get?
On the non-working machine, I can't install it.
On the working machine I encounter no issues.
-
Maybe it's something with yum-cron as that is installed on this VM, and is set to apply updates. Whereas the other VM has yum cron, but isn't set to apply the updates.
Still doesn't really explain why ping wouldn't work to the internet.
-
What does this look like on the non-working CentOS?
/etc/sysconfig/network-scripts/ifcfg-eth0
is onboot enabled?
Edit: nvm. I see you can ping locally and to the gateway
-
-
This is what I get from the non-working VM.
yum update Loaded plugins: fastestmirror Loading mirror speeds from cached hostfile * base: mirror.atlanticmetro.net * extras: bay.uchicago.edu * updates: centos.s.uw.edu http://repos-va.psychz.net/centos/7.6.1810/os/x86_64/repodata/repomd.xml: [Errno 14] curl#6 - "Could not resolve host: repos-va.psychz.net; Unknown error" Trying other mirror.It just keeps going through all of the mirrors, obviously.
I've disabled and turned off yum-cron and have reboot numerous times. So that isn't it. I can ping locally by IP and DNS from the non-working machine.
-
I guess I'll just restore the VM and see what happens from it.
-
@DustinB3403 said in Troubleshooting Help Requested:
This is what I get from the non-working VM.
yum update
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile- base: mirror.atlanticmetro.net
- extras: bay.uchicago.edu
- updates: centos.s.uw.edu
http://repos-va.psychz.net/centos/7.6.1810/os/x86_64/repodata/repomd.xml: [Errno 14] curl#6 - "Could not resolve host: repos-va.psychz.net; Unknown error"
Trying other mirror.
It just keeps going through all of the mirrors, obviously.
I've disabled and turned off yum-cron and have reboot numerous times. So that isn't it. I can ping locally by IP and DNS from the non-working machine.
Did I see right that
tracerouteisn't installed? -
@travisdh1 yes.
-
And it isn't required either, as this system has updated fine for months.
-
@DustinB3403 said in Troubleshooting Help Requested:
@travisdh1 yes.
Makes it difficult to see weather there are peering-point issues between you and the mirror sites.
edit: Yes, I've seen networks block distribution mirror sites before. So much fun trying to get those fixed.