| Title: | DEC TCP/IP Services for OpenVMS |
| Notice: | Note 2-SSB Kits, 3-FT Kits, 4-Patch Info, 7-QAR System |
| Moderator: | ucxaxp.ucx.lkg.dec.com::TIBBERT |
| Created: | Thu Nov 17 1994 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 5568 |
| Total number of notes: | 21492 |
Is it normal that the secondary DNS-server should stop working in DNS
if the Primary DNS-server is gone when it wants an update?
To be more specific:
A customer have node A working as a primary DNS-server and node B as a
secondary DNS-server. They decide to upgrade UCX on node A and stop
UCX on this node. Node B recognice this and DNS is working fine in the
network. When the upgrade of UCX on node A is finished they didn't
realize that BIND wasn't enabled on this node anymore, still everything
worked ok in the net. Two days later (we have looked in the logs) node
B wants to be updated from the primary DNS-server. This messages
continues at least 12 hours and after 16 hours when they arrive in the
morning DNS isn't working anymore in their network. They now found out the
BIND-mistake in node A, fix it and all is working fine again.
The only thing left is the question, was this expected or what is mis-
configured? Node A was running UCX V4.1 ECO2 before the upgrade and
node B is running UCX V4.1 ECO5 (bugfixed ECO4 prerelease).
Is it better to have more than one secondary server ?
/Dan Lundeborg TSC Sweden
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 5547.1 | normal behaviour | UTRTSC::KNOPPERS | Oswald Knoppers | Thu May 29 1997 05:23 | 23 |
> The only thing left is the question, was this expected or what is mis- > configured? Node A was running UCX V4.1 ECO2 before the upgrade and > node B is running UCX V4.1 ECO5 (bugfixed ECO4 prerelease). Yes this is normal. This is set with the 'expire' field in the SOA record. If the secondary server fails to contact the primary for 'expire' seconds it will stop servicing requests. > Is it better to have more than one secondary server ? Not for this problem, other secondary servers will stop servicing requests just the same. I always configure the secondary servers' resolver pointing to itself. This ensures that the bind process is started on the secondary server. If you don't do this, the possibility exist that this process is never created while the primary server runs. If the primary server goes down, clients start using the secondary server but at this time it's data is already expired. Regards, Oswald | |||||
| 5547.2 | Thanks | RULLE::LUNDEBORG | Fri May 30 1997 03:30 | 7 | |
Thanks for your reply, the customer bought your secondary server
solution.
/Dan Lundeborg TDC Sweden
| |||||