Transfers
Jim Reid
jim at rfc1035.com
Sat Apr 14 08:07:10 UTC 2001
>>>>> "Craig" == hibbert, craig <hibbert_craig at emc.com> writes:
Craig> I am having some problems with transfers from our primary
Craig> server. Both the primary and secondary servers are running
Craig> 8.2.3 and everything works ok except the zone transfers. I
Craig> would expect that when the refresh time is up on any
Craig> particular zone, that zone would update from the
Craig> primary. The only way that works in our environment is to
Craig> Kill -1 (or HUP) the named process running on the
Craig> primary. This appears to be re-read into memory by the
Craig> primary and then a zone transfer takes place
Craig> successfully. My question is... can the same results be
Craig> obtained without killing the process on the primary every
Craig> time we need to do a transfer?
Zone transfers are performed by slave (secondary) servers. Anything
you do to the master server has no bearing on how the slaves perform
transfers. Reloading the master server -- use ndc or rndc, don't send
signals! -- will make no difference to what the slave servers do. They
will only perform a zone transfer and load a new copy of a zone if
they see that the master server's copy of the zone has a higher SOA
serial number than they do. Those transfers will only happen if the
master server allows them to and is authoritative for the zone and
there's no connectivity problem between the servers.
You've provided no useful information about your problem, like the
logs from the master and slave server, the names of the zones and the
servers in question. Post these details if you want help. Show these
files as-is. Don't prettify them or sanitise them. What you post
should be exactly what your name server sees or prints.
More information about the bind-users
mailing list