[Dnsmasq-discuss] --all-servers option

Petr Menšík pemensik at redhat.com
Thu May 28 21:21:35 BST 2020


Hello David,

first of all, RHEL 6 version is quite old, I would say almost ancient.
But I understand the distribution still has it. I would suggest
upgrading to RHEL 7 at least, better to version 8.

Please try it with a more recent version. It is quite possible it is
fixed for several years. It would not be fixed on RHEL 6.

On 5/27/20 10:37 AM, David Krupička wrote:
> Dear all,
> 
> I am using version
> Name        : dnsmasq
> Arch        : x86_64
> Version     : 2.48
> Release     : 18.el6_9
> 
> I have configured this (domains and IP's are fake)
> # Add other name servers here, with domain specs if they are for
> # non-public domains.
> #server=/localnet/192.168.0.1
> all-servers
> server=/domain1.internal/172.20.11.10    # THIS server does not exist,
> simulating like it was down
> server=/domain1.internal/172.20.11.11
> server=/domain2.internal/172.21.11.12
> server=/domain2.internal/172.21.11.13
> 
> Regardless of running as a daemon or not, the all-servers option does not
> work as I would expect.
I think you forgot to explain what is your expected behavior. What do
you think it should do?
> Reading of /etc/resolv.conf is disabled,strict-order too. I need multiple
> servers per domain.
> 
> Was this ever dealt with in higher versions? Could it be a bug or it simply
> does not work like this?
> There is 5 seconds delay, see below... Can I somehow reduce it? Or what
> should I do?
I think similar bug was recently fixed on RHEL 7. It should try both
servers, but haven't.
> 
> Example:
> root at hostx:~# date && nslookup server1.domain1.internal && date
> Wed May 27 10:22:05 CEST 2020
> Server:         127.0.0.1
> Address:        127.0.0.1#53
> 
> Name:   server1.domain1.internal
> Address: 172.21.100.33
> 
> Wed May 27 10:22:10 CEST 2020
> 
> Log:
> dnsmasq: query[A] server1.domain1.internal from 127.0.0.1
> dnsmasq: forwarded server1.domain1.internal to 172.20.11.10
> dnsmasq: query[A] server1.domain1.internal from 127.0.0.1
> dnsmasq: forwarded server1.domain1.internal to 172.20.11.11
> dnsmasq: forwarded server1.domain1.internal to 172.20.11.10
> dnsmasq: reply server1.domain1.internal is 172.21.100.33
> 
> root at iumdbtestflo01:~# tcpdump -ni any port 53
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on any, link-type LINUX_SLL (Linux cooked), capture size 65535
> bytes
> 10:30:55.813271 IP 127.0.0.1.9772 > 127.0.0.1.domain: 61442+ A?
> server1.domain1.internal. (61)
> 10:30:55.819510 IP 172.17.211.6.sun-sr-https > 172.20.11.10.domain: 36519+
> A?  server1.domain1.internal.   (61)
> 10:31:00.808765 IP 127.0.0.1.9772 > 127.0.0.1.domain: 61442+ A?
> server1.domain1.internal.   (61)
> 10:31:00.808961 IP 172.17.211.6.sun-sr-https > 172.20.11.11.domain: 36519+
> A?  server1.domain1.internal.   (61)
> 10:31:00.809006 IP 172.17.211.6.sun-sr-https > 172.20.11.10.domain: 36519+
> A?  server1.domain1.internal.   (61)
> 10:31:00.811297 IP 172.20.11.11.domain > 172.17.211.6.sun-sr-https: 36519*
> 1/1/1 A 172.21.100.33 (112)
> 10:31:00.811412 IP 127.0.0.1.domain > 127.0.0.1.9772: 61442* 1/1/1 A
> 172.21.100.33 (112)
> 
> Best regards
> David

Just for my curiosity, why do you want both servers to be used? Default
configuration tries random server first and retries only when it does
not respond. Would it work for you that way?

Best regards,
Petr

-- 
Petr Menšík
Software Engineer
Red Hat, http://www.redhat.com/
email: pemensik at redhat.com
PGP: DFCF908DB7C87E8E529925BC4931CA5B6C9FC5CB




More information about the Dnsmasq-discuss mailing list