All sites have stop resolving!

13 posts / 0 new
Last post
#1 Mon, 04/15/2019 - 15:14
Rockyuk

All sites have stop resolving!

Hi Everyone, I finally got my server up and running with Ubuntu 18.04 and Virtualmin. Everything was working fine for about two to three days. Then this morning nothing none of the sites resolve anymore? Checked all the dns settings and everything seems to be fine nothing has changed!

Does anyone have any ideas of what could have stopped the resolution suddenly like that?

Thanks

Rockyuk

Mon, 04/15/2019 - 20:47
andreychek

Howdy,

Hmm, is BIND running?

If BIND has stopped for some reason, and your server is the nameserver for your domains, that could cause similar issues.

If your server is the nameserver, you may also want to ensure that 127.0.0.1 is listed in /etc/resolv.conf.

-Eric

Tue, 04/16/2019 - 03:26
Jfro

Is system rebooted in that time? ( or network restarted)

While the resolv.conf with some hosters are replaced then for default ones, i am just testing here and having this on openstack instance.

If so make it right and lock with chattr +i resolv.conf

to unlock chattr -i resolv.conf

Tue, 04/16/2019 - 09:44
tzohosting

I have reported a similar issue a month ago using AWS EC2 in a VPC, I have yet to get it resolved. I am beginning to think the Virtualmin may not be the best option with Ubuntu 18.04 (official AWS AMI) if on AWS EC2.

My guess is your server rebooted and lost all the internal IPs and has defaulted to a single IP. This has happened on multiple instances on fresh installs of Virtualmin. Never had this issue on Ubuntu 14 or 16. See issue here...

https://www.virtualmin.com/node/65079

if anyone else has anything thoughts, I am all ears.

Tue, 04/16/2019 - 09:58 (Reply to #4)
Jfro

Sorry your link has no access

Access Denied You are not authorized to access this page.

Are there things like cloudinit on those? https://cloudinit.readthedocs.io/en/latest/topics/network-config.html#ne...

and differences ubuntu

Netplan is the replacement of ifupdown in Ubuntu. It is a new way to configure your network. However we found that it is incompatible with dhcpv6. This would make ipv6 non functional on this image when dhcp ipv6

also eno/ens network names. and not eth...

Tue, 04/16/2019 - 10:02 (Reply to #5)
tzohosting

Sorry, I have a paid license and may only be for paying customers? In any case, yes on cloudinit. Here is the gist of the issue I reported.

"The specific issue is around creating virtual servers and assigning private IPs. Once you reboot, all of the IPs get lost and Virtualmin complains that the server IP has changed.

I have done a fresh install of Ubuntu 18.04.2 (Official AWS Provided AMI) and Set a FQDN hostname and installed latest Virtualmin under VPC. Install of Virtualmin was error free.

I can create virtual servers fine and give them private IPs and assign public IPs via AWS console. This works and serves sites fine. The network interface in Webmin > Networking > Network Configuration > Network and the interface all look correct.

However, when a reboot occurs, the network interface loses all the IPs for the virtual servers and defaults to the incorrect IP. It actually changes and defaults to the highest IP. Example, the server IP is 10.0.13.10, all virtual servers will have their own private IP as 10.0.13.11, 10.0.13.12, etc... But once a reboot happens, all virtual server IPs are lost and defaults to 10.0.13.12.

I have no issue with this on Ubuntu 16.04.

My questions is, is this still a known issue? Is there a configuration change that must take place to work with multiple IPs?"

Tue, 04/16/2019 - 10:10
Jfro

You see some important diferences so it could be working if you change / revert the not working parts to as it was in old version. I have edited my post above

Also some networking parts are replaced / overwriten at reboots as often resolv.conf .... cloudinit for cloud / openstack and so on.

Example of a openstack where > https://www.cloudvps.com/helpcenter/openstack-faq/openstack-images/ubunt...

I'm testing also private ips AND CENTOS ( so no experience with UBUNTU ! IPV4 working ipv6 for the moment can't get it running for outside access, tough the ipv6 external is on the instance as i can see in network for eth1

Thu, 05/09/2019 - 14:29
sz00gun

Still have the same problem, any solutions so far? Did anyone make the webmin working with Ubuntu 18.04 with couple of IP?

Thu, 05/09/2019 - 15:00 (Reply to #8)
tzohosting

Assuming a reboot caused this and you are on a cloud server? Not that I have found, I also have a paid license and my support issue has still gone unanswered. Virtualmin just does not work, out of the box, with Ubuntu 18.04/Netplan, on AWS EC2 instances. Reboots will overwrite all changes Virtualmin adds to your network interface.

Thu, 05/09/2019 - 15:27
sz00gun

Indeed, I have exactly the same problems, sounds like I have to say good bye to webmin, however I don't want it.... Now doing tests with Centos Web panel...

Thu, 05/09/2019 - 16:24
Jfro

on AWS EC2 instances. Reboots will overwrite all changes

Those things are also happening with openstack instances. ( cloudinit responsibel) You have to protect the files for those changes, and important to have them updated manually wen needed. ( take care with all that dhcp stuff)

That has nothing to do with virtualmin itself is common for instances in cloud plattforms. resolv.conf you have to protect and so on.

Or you should have more right or support with that platform hoster / admins if something is possible. Here with openstack is answer for example the resolv.conf a NO, solved it with chatr https://www.tecmint.com/chattr-command-examples/

Here on CENTOS 7 for now it seems ok with more ip's ( internal private ip's, with table to public ip's)

Thu, 05/09/2019 - 16:46 (Reply to #11)
tzohosting

Thanks for providing the external link. Will review. I would assume that if Virtualmin is going to manage the network interface for virtual hosts (and it does), that it would also account for cloudinit as well.

Thu, 05/09/2019 - 18:00
Jfro

Not sure and no time , we only have those reboot changes problem with resolv.conf in openstack instances centos7 and virtualmin sofar as i can see with simple firsts testing.

So differences with my situation are: CENTOS 7 OPENSTACK. Other Hoster. No payd license virtualmin and i'must be crazy wanting to have instances on openstack with control panel ( as virtualmin or whatever) i presume. ;)

Forgot most important difference: I'm TOTAL ANTI AMAZON ! ;)