1

Our network has a number of systems that are dynamically allocated IPs (which happen to be in the public range). These systems also have a local hostname, which is making it into the DNS resolver cache with whatever IP the system currently has. When the system gets a new IP, clients trying to connect via hostname are having to manually flush the DNS resolver cache before they can find the host at its new IP.

I believe that simply preventing the cache from including hosts within a certain IP range should resolve the issue. Working the other way, if it's possible to exclude non-fully-qualified hostnames that should also resolve the issue.

I've done quite a bit of Googling, but nothing useful is showing up.

1 Answer 1

0

The issue only comes up when the IP address changes, correct? If that's the case, you can set the TTL on the individual DNS records to 1 second. That effectively prevents just those records from being cached.

You can also disable the DNS client service on workstations. That prevents the cache from working also, but that affects all DNS lookups, not just the ones you're interested in.

2
  • How exactly would I set the TTL for a range of IPs on the client? Keep in mind, I'm looking for a client-centric solution. Don't have an easy to change how the powers that be are handling the DNS server.
    – Dan
    Sep 18, 2014 at 19:07
  • Ah. In that case you can't modify it on a per-record basis. You can disable caching entirely by stopping the DNS Client service. Despite it's name, it does not disable DNS. Only the caching of DNS records, forcing the client to query for every DNS lookup every time. Alternatively, you can follow this KB article from Micrsoft to override the TTL value on the client. This eases up the load on your DNS server a bit. But it's still a global value. You still can't set it on a per-record basis.
    – Wes Sayeed
    Sep 18, 2014 at 19:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .