Win2k3 Production Servers DNS records become stale and get deleted Roch 11/12/2008 5:08:02 PM Hi there, I wonder if somebody can point me in the right direction. We have around 1500 users and aprox. 200 servers. For several weeks now we have received calls from users stating that they can't access certain server. After checking, we have found out that the dns record has been deleted. I know that I can uncheck the box in the DNS console for "delete this record when it becomes stale" and that would fix the problem at least with the most important records but that would be cumbersome to I am trying to figure out the process. I know that there are three DNS elements involved in the process: scavenging, no-refresh interval, and refresh interval. Our no-refresh and refresh intervals are set to 1 day. Scavenging on the two main Enterprise DNS servers are set to 1 day and 7 days. DHCP lease is set for all our scopes to 8 days. The way I understand the process works is that no-refresh is the time interval that DNS doesn't allow refreshing DNS information (this would help limiting dns traffic in some environments). After no-refresh interval is over, no-refresh interval kicks in and the devices are required to register with dns. They do that via dhcp client service, restarting net logon service, or typing ipconfig/registerdns. If at the end of the refresh interval the device doesn't register, the record becomes stale and scavenging kicks in and deletes the record. If I am not mistaking the recommendation is to set refresh interval close to the DHCP lease expiration but I don't fully understand why. I am interested in knowing what is the recommended refresh and no-refresh interval settings for an enterprise like ours and I would like to learn from you about some good recommendations. Thank you in advance, Roch

RE: Win2k3 Production Servers DNS records become stale and get deleted James Yeomans BSc, MCSE 11/12/2008 7:36:13 PM Hi there, best practice is usually to leave the no-refresh and refresh intervals to 7 days each. 1 day will result in to much scavenging activity and records may be deleted before they are refreshed. Do your servers have static ip addresses, they should not be configured by DHCP. Servers configured with static ip addresses will refresh their dns records often enough (every 24 hours i believe) to negate the possibility of being scavenged assuming the scavenging period is set appropriately, i.e. not less than the DHCP lease time. This so that machines with dynamic addresses are not at risk of having their record deleted from DNS. Hope that helps James. -- James Yeomans, BSc, MCSE Ask me directly at: http://www.justaskjames.co.uk "Roch" wrote: [Quoted Text] > Hi there, > I wonder if somebody can point me in the right direction. We have around > 1500 users and aprox. 200 servers. For several weeks now we have received > calls from users stating that they can't access certain server. After > checking, we have found out that the dns record has been deleted. > I know that I can uncheck the box in the DNS console for "delete this record > when it becomes stale" and that would fix the problem at least with the most > important records but that would be cumbersome to I am trying to figure out > the process. > I know that there are three DNS elements involved in the process: > scavenging, no-refresh interval, and refresh interval. > Our no-refresh and refresh intervals are set to 1 day. Scavenging on the two > main Enterprise DNS servers are set to 1 day and 7 days. DHCP lease is set > for all our scopes to 8 days. > The way I understand the process works is that no-refresh is the time > interval that DNS doesn't allow refreshing DNS information (this would help > limiting dns traffic in some environments). After no-refresh interval is > over, no-refresh interval kicks in and the devices are required to register > with dns. They do that via dhcp client service, restarting net logon service, > or typing ipconfig/registerdns. If at the end of the refresh interval the > device doesn't register, the record becomes stale and scavenging kicks in and > deletes the record. > If I am not mistaking the recommendation is to set refresh interval close to > the DHCP lease expiration but I don't fully understand why. > > I am interested in knowing what is the recommended refresh and no-refresh > interval settings for an enterprise like ours and I would like to learn from > you about some good recommendations. > > Thank you in advance, > > Roch

RE: Win2k3 Production Servers DNS records become stale and get del Roch 11/12/2008 9:20:12 PM Thanks James for the quick response your posting was very helpful. Do you know if for an enterprise of our size would be recommended to have only one DNS performing the scavenging instead?

RE: Win2k3 Production Servers DNS records become stale and get del James Yeomans BSc, MCSE 11/12/2008 9:58:03 PM Yes you only need 1 dns server doing the scavenging as the records will not actually be removed until the necessary replication between dns servers has taken place. James. -- James Yeomans, BSc, MCSE Ask me directly at: http://www.justaskjames.co.uk "Roch" wrote: [Quoted Text] > Thanks James for the quick response your posting was very helpful. > Do you know if for an enterprise of our size would be recommended to have > only one DNS performing the scavenging instead? > >

RE: Win2k3 Production Servers DNS records become stale and get del Roch 11/12/2008 10:22:05 PM Thanks James. I wasn't sure so I appreciate your help. "James Yeomans BSc, MCSE" wrote: [Quoted Text] > Yes you only need 1 dns server doing the scavenging as the records will not > actually be removed until the necessary replication between dns servers has > taken place. > James. > -- > James Yeomans, BSc, MCSE > Ask me directly at: http://www.justaskjames.co.uk > > > "Roch" wrote: > > > Thanks James for the quick response your posting was very helpful. > > Do you know if for an enterprise of our size would be recommended to have > > only one DNS performing the scavenging instead? > > > >