[Samba] dns replication error due to deleted records
luca at wetron.es
Sun Oct 13 13:44:39 UTC 2019
El 11/10/19 a les 17:29, Rowland penny via samba ha escrit:
>> it's here:
> Your main problem appears to be that you have a lot of duplicate objects
> in the DNSDOMAIN context, these are the records with '0ACNF' in them.
I think I know what happened: a while ago we had a problem with our
virtualization infrastructure that lead to a partial failure of dc2 and
a catastrophic failure of dc1 (not to mention failures in every other
server). I could recover dc2 pretty quickly, but I had to recover dc1
That's *not* the problem though: while I was recovering dc1, dns
resolution was unbearably slow on every server, because I had dc1 as the
first nameserver and dc2 as the second.
To avoid that happening again in the future, I installed dnsmasq (since
it caches results and it's quicker to detect that a nameserver is down)
on every server, including the dhcp server.
Now, the script on the dhcp server that updates the dns record checks if
the record exists using "host", but now it gets its result from dnsmasq,
hence probably causing those duplicates.
I amended the script to check directly on dc1, lets' see if it fixes the
issue (a issue that shouldn't happen anyway: I'm using the published
interfaces to update the records, so that shouldn't stop the replication).
Wetron Automation Technology http://www.wetron.es/
Tel. +34 93 5883004 (Ext.3010) Fax +34 93 5883007
More information about the samba