DNS / NTP Server on HyperVisor not set


Userlevel 2
Badge +14
I have installed a new AHV cluster - but i keep getting the DNS / NTP error on the host

I investigated and it seems like DNS / NTP Service is not setup from the cluster setup add-to etc.

What can go wrong here? The resolv.conf file is missing on 2/3 AHV hosts - this seems pretty wierd..

This topic has been closed for comments

19 replies

Userlevel 2
Badge +14
I have tried ncli cluster remove-from-name-servers servers=8.8.8.8,8.8.4.4 and re-add to the servers - but the resolv.conf file is not recreated on the hypervisor..
Userlevel 2
Badge +14
When i check one of the hosts the resolv.conf file is created like this

# Auto generated by HypervisorDnsConfig on Tue Mar 14 11:12:39 2017
nameserver 8.8.8.8nameserver 8.8.4.4

But it does not exist on the other hosts - i have tried rebooting the whole cluster and reset it but its only set on one of the hosts..
Userlevel 2
Badge +14
jrack or Jon - this seems like its something with the newest image maybe?

Have you seen this problem before?
Userlevel 7
Badge +25
I don't use Google's DNS so not the greatest tester. On all of my AHV nodes I show my configured DNS though so not sure what is up with yours.

Try and create it manually for giggles?
Userlevel 2
Badge +14
Tried create the resolv.conf by hand

Have you installed from newest image or upgraded from old?
Userlevel 7
Badge +25
All upgrades over the past couple of years and not even at the latest so yeah def could be something in the latest.
Badge +5
I am having the same problem with my most recent update.
Userlevel 2
Badge +14
Okay - waiting for Jon - he may know about this issue.. reinstalled the cluster was the same issue.. 😞 Also on a CLEAN install just created the cluster
Userlevel 2
Badge +14
Is it all your Hypervisors or only some of them?

And are you able to etc. ping www.google.com?

I did try to set the resolv.conf by hand but did not work.. Seems like its more broken than just the file missing..
Badge +5
I just installed a new 3 node cluster using the latest full stable install. DNS is not configured on any of the hosts. it is configured on the CVMs. I think that the physical network connection on the hosts were not actually working when I did the install, perhaps that caused this?

[root@NTNX-7c151a10-A ~]# ping www.google.comping: www.google.com: Name or service not known[root@NTNX-7c151a10-A ~]# ping 10.10.12.1PING 10.10.12.1 (10.10.12.1) 56(84) bytes of data.64 bytes from 10.10.12.1: icmp_seq=1 ttl=255 time=0.829 ms64 bytes from 10.10.12.1: icmp_seq=2 ttl=255 time=0.535 ms^C

Is there a command to add it to the host?
Userlevel 2
Badge +14
It must be an issue - i have tried again works with the old build but the hypervisor is not supporting my 10g interfaces .. On the new build i keep getting the issue with the DNS/NTP - and it does not propergate to the Hypervisort from the CVM

Jon - have you heard about this?
Badge +2
Can confirm this bug. When I was creating the cluster yesterday same error message "Couldn't set DNS/NTP on hypervisor" and no resolv.conf files available.

Workaround: Copy resolv.conf from CVM to Hypervisor-Node than recheck with
ncc health_checks system_checks dns_server_checkand
ping google.de
Badge +1
I'm having this same issue on a new clean install of the 20180123.170 code as well.
Userlevel 2
Badge +1
+1

I am having the same issue in 2018.01.31...Will setting the time zone help?
Badge +1
manually creating resolv.conf then running ntpdate on each host resolved this.
Userlevel 2
Badge +1
Do you have a link to an article that could walk me through how to do that?


manually creating resolv.conf then running ntpdate on each host resolved this.
Userlevel 2
Badge +1
I tried the following and I think it worked...:

Set the timezone on each AHV server:
hostssh "date; mv /etc/localtime /etc/localtime.bak; ln -s /usr/share/zoneinfo/America/Boise /etc/localtime; date"
**Note: Make sure you change the time zone to meet your needs

Set the timezone on the cluster:
ncli cluster set-timezone timezone=America/Boise

Connect to the AHV host:
ssh root@192.168.5.1

Create/Open resolv.conf:
vi /etc/resolv.conf

Add the following text into resolv.conf:
# Auto generated by DnsConfig on Thu Mar 1 14:38:17 2018

nameserver pool.ntp.org
**Note: Change the NTP server to meet your needs

Restart the Network on the Host:
service network restart

Update NTP:
service ntpd stop
ntpdate pool.ntp.org
service ntpd start

You have to wait a few minutes before everything gets sorted out. I was able to get one of my three nodes to work correctly. Unfortunately I didn't document the process...hopefully the above works.

Are there any steps that I missed?
Userlevel 2
Badge +1
I am happy to report that those steps worked for me. All three of my nodes are working as expected.

I tried the following and I think it worked...:

Set the timezone on each AHV server:
hostssh "date; mv /etc/localtime /etc/localtime.bak; ln -s /usr/share/zoneinfo/America/Boise /etc/localtime; date"
**Note: Make sure you change the time zone to meet your needs

Set the timezone on the cluster:
ncli cluster set-timezone timezone=America/Boise

Connect to the AHV host:
ssh root@192.168.5.1

Create/Open resolv.conf:
vi /etc/resolv.conf

Add the following text into resolv.conf:
# Auto generated by DnsConfig on Thu Mar 1 14:38:17 2018

nameserver pool.ntp.org
**Note: Change the NTP server to meet your needs

Restart the Network on the Host:
service network restart

Update NTP:
service ntpd stop
ntpdate pool.ntp.org
service ntpd start

You have to wait a few minutes before everything gets sorted out. I was able to get one of my three nodes to work correctly. Unfortunately I didn't document the process...hopefully the above works.

Are there any steps that I missed?
Badge +8
I tried the following and I think it worked...:

Set the timezone on each AHV server:
hostssh "date; mv /etc/localtime /etc/localtime.bak; ln -s /usr/share/zoneinfo/America/Boise /etc/localtime; date"
**Note: Make sure you change the time zone to meet your needs

Set the timezone on the cluster:
ncli cluster set-timezone timezone=America/Boise

Connect to the AHV host:
ssh root@192.168.5.1

Create/Open resolv.conf:
vi /etc/resolv.conf

Add the following text into resolv.conf:
# Auto generated by DnsConfig on Thu Mar 1 14:38:17 2018

nameserver pool.ntp.org
**Note: Change the NTP server to meet your needs

Restart the Network on the Host:
service network restart

Update NTP:
service ntpd stop
ntpdate pool.ntp.org
service ntpd start

You have to wait a few minutes before everything gets sorted out. I was able to get one of my three nodes to work correctly. Unfortunately I didn't document the process...hopefully the above works.

Are there any steps that I missed?

Seems that no matter what I do, at least one of my nodes wont stay 'fixed' with this fix while the other two are, for now. Also, editing that file seems like a silly step since its overridden by whatever is configured in Prism -- the change doesnt stay. And in prism one must define nameservers by IP not FQDN. My resolv.conf also says autogenerated by HypervisorDnsConfig not just DnsConfig

My version is 2017.07.20