VDS Can't Resolve Domains

[Runamuk]

My VDS setup was working perfectly up until yesterday. Nothing was really changed and I can't figure out why it's failing.

Here is the setup.

VDS running Ubuntu 10.10. Runs GPC with Urban Terror servers and B3 and some custom scripts. Anything requiring web access fails to resolve.

Example:
Urban Terror console gives failed to resolve to master.urbanterror.info. As well as master2 and master.qtracker.com. They all fail to resolve. The server takes longer to load but eventually loads and will crash shortly after.

B3 fails to check for an update with a resolve error. Our b3 bots cannot post on our website due to resolution errors.

Custom script gives this error:
ERROR: unable to access VSP dbase(this database is hosted on a website).
Map Cycled, Pausing a cycle

Attempted multiple VDS reboots and disabling of firewall. Nothing has really been changed other than a system downgrade.

I also manage another VDS and it has the exact same issue.

Not sure where to start looking for this issue.
West Senior

[Runamuk]

So...as soon as I posted this it hit me to check the resolver file.

I changed the IP(Not sure what the included IP's were for or how they got there) to 8.8.8.8 and now everything resolves fine. Not exactly sure why this issue just sprung up but this fixed it.

/etc/resolv.conf

[IcEWoLF]

http://www.linuxquestions.org/questions ... ns-888204/

nvm, you found your answer.

[Edge100x]

This likely relates to the issue we mentioned on Facebook/Twitter on 3/28:

If you run an unmanaged VDS or dedicated machine and you're having problems with DNS resolution today, check to make sure that your DNS servers are set to 8.8.8.8 and 8.8.4.4 (Google's). Some older machines here were originally set up to use the UW servers (128.95.120.1 and 128.95.112.1), and those seem to be broken today.

[Runamuk]

This likely relates to the issue we mentioned on Facebook/Twitter on 3/28:

If you run an unmanaged VDS or dedicated machine and you're having problems with DNS resolution today, check to make sure that your DNS servers are set to 8.8.8.8 and 8.8.4.4 (Google's). Some older machines here were originally set up to use the UW servers (128.95.120.1 and 128.95.112.1), and those seem to be broken today.

Yes, well I apparently missed that post on Facebook. Seems like that would have made a good event message in the control panel or forum post? Facebook isn't the best place at making sure you get the info. I like your page but don't visit it everyday looking for updates. If they show in my news feed I see them. I didn't send in a support ticket because the usual answer is that it's an unmanaged VDS so we don't support them, give the forums a try. Good to know it wasn't just me...

[Edge100x]

The forums are also a great place to ask this, I agree. In a support request, we would have also described the solution (the post to Facebook was mostly in response to several support requests that we'd received). I initially investigated the issue as a possible network/bandwidth problem, and we definitely support those directly, along with hardware/Xen issues.

I hesitate to post to the news on the site or to push out to everyone's events log when the issue does not apply to many customers. Partially this is because I know that the events log sends out emails to most customers, and I know that many might consider this to be spam. Facebook is a bit different because it's opt-in and "pull" instead of "push".

If a situation like this occurs again, I may write a script to attempt to better target who the message would apply to, and post to the "Events log" of just those customers.