Süsteemihaldus 2021/22 kevad

Overview of Lab 4 :

Welcome to the DNS lab. Here is the short action list we will do in this lab:

1. Verifying previous lab tasks
2. Installing updates
3. Opening firewalls
4. Installing Domain Name Server
5. Configuring own local resolver server (allowing recursion and caching)
6. Configuring own domain zone, understanding "world" and "local" resolvers

1. Make sure you did finish all the tasks of the previous week.

PS! Before continuing, please make sure you have completed all previous labs.

Additionally - Securing your ssh server

One thing that makes the SSH server more secure is disabling password authentication. We disable it since we use our private key to log in, and we disable password login because people often use bad (easy to guess) passwords. We want to eliminate that risk, and since we use a key, we disable it.

2. Check for the updates using CentOS package manager

We have worked with package managers yum or dnf during Lab2.

After upgrading the system, it is wise to clean the left-overs: the downloaded "rpm" packages. Moreover, it is wise to remove the packages that are not used anymore. They are two commands responsible for performing these actions:

# Uninstall the not used packages
yum autoremove -y
yum clean all

Extra reading material for package managers

DNF Command Reference
YUM cheat sheet

3. Firewall - Open UDP and TCP port 53 for the lab

Computers and especially servers usually utilize their own firewalls for security reasons. This allows the most privileged user of the machine to dictate how the network should go to/through/from the machine. The most usual use for this is opening ports for services, so they could be reached from the outside world.

Usually, to top it off, with any sensible cloud provider you have their extra (second) layer of firewall, which has to be configured from outside the machine. This is to make sure, that even if your machine becomes compromised, network security policies (firewall rules) could not be changed easily.

From this lab onward we need UDP and TCP port 53 to be open for our VMs so that we could build a working DNS solution. DNS usually uses port 53/UDP for everything, but with big enough queries it switches over to 53/TCP. We also need to open port 8301 both UDP/TCP for our scoring to be able to a service discovery tool called @Consul@ pre-installed by us on your machines.

To do that for the cloud provider layer, follow these steps:

When you open any other port, you also have to do it in the ETAIS interface. This is not a one-time thing.

Let's make sure the DNS service can also be accessed on the host level:

The output of previous commands should be a success.

4. Disable IPv6

The ETAIS cloud does not suggest any IPv6 network providers; therefore, it is pointless to keep the IPv6 stack up in our VMs. It might also cause problems in some particular instances. In this section, we will disable the IPv6 stack by alternating kernel runtime parameters.

System-wide configuration variables are managed using sysctl utility. Example:

• To print the whole list of system settings, use:
  • # sysctl -A
• To see specific variable value:
  • # sysctl variableName
• To set a specific variable to a new value use:
  • # sysctl <variable>=<value>

However, this will change only running system settings and reset default values after a system reboot. To alter the default values put the corresponding variables at the end of /etc/sysctl.conf file, for example, the content of /etc/sysctl.conf (only a few last lines shown):

...

 # My default value of aa.bb.cc_var1 variable
 # NB! This is an example, do not use this!
 aa.bb.cc_var1=1

...

We do not configure any service to rely on IPv6; hence we do not need the system-wide IPv6 stack. Let's disable it using well studied sysctl utility.

Now, if you check your network interfaces using the ip addr l command, there should be no inet6 prefixed lines.

5. Domain Name System (DNS) using BIND software

DNS (Domain Name Service) is the primary service used on the Internet. Domain name services translate names meaningful to humans into numerical addresses (and vice versa). Practically every system relies on a working domain name service. The domain name service does not have to be set up in-house; the Internet Service Provider (ISP) often provides it.

Extra materials

• Wikipedia: Domain Name System

Domain Hierarchy for System Administration Labs

We will use our own domain hierarchy for our practicals, starting from the top-level domains. DNS management is based on strict delegation, so we cannot introduce our "fake" top-level domains into the public DNS system used on the Internet. Everything should be working inside the University network or VPN.

The domain hierarchy is divided into zones based on authority:

• The authority over root domain and top-level domains will belong to the teaching staff. We will be using top-level domain .sa.cs.ut.ee.
• Every participant/system will have authority over a second-level domain vm_name.sa.cs.ut.ee, under which one or more names may be defined. "vm_name" refers to the name to the machine you've set inside minu.etais.ee. It is mandatory to have a name for every host/subsystem in the lab group under these second-level domains.
• The general form for these names is <service>.vm_name.sa.cs.ut.ee Here "<service>" refers to any services we will be setting up in the future.

For example:

• If student John Doe has his UT study number book B12345 and his VM was named B12345 then full hostname is B12345.sa.cs.ut.ee. You will have a different matrix number representing your VM name (host name), but the top-level domain .sa.cs.ut.ee will stay the same in your FQDN.

Changing hostname

By default, all VM instances created in this course are given a hostname of <vm_name>.novalocal by ETAIS during the initial setup. During this lab we are going to set up a personal DNS server (serving personal domain <vm_name>.sa.cs.ut.ee). Therefore we have to prepare the VM by changing its hostname from the ETAIS' default notation to full qualified domain notation, for example, for "B12345" in "B12345.sa.cs.ut.ee" domain:

* short notation: B12345
* fqdn notation: B12345.sa.cs.ut.ee

The hostname is held in file /etc/hostname. By altering the single line, you will change the name of your host. In addition, there is a /etc/hosts file, which provides fixed host name to ip address mappings, and has the following content (only the line we are interested in is shown):

127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
... 

The line shown above allows resolving hostname localhost to a local IP address 127.0.0.1 (without any requests to external services like DNS). This is especially useful when testing your services locally, if you consider the simple command, like:

$ ping localhost
PING localhost (127.0.0.1) 56(84) bytes of data.
64 bytes from localhost (127.0.0.1): icmp_seq=1 ttl=64 time=0.052 ms
64 bytes from localhost (127.0.0.1): icmp_seq=2 ttl=64 time=0.060 ms
...

We need to add another line in /etc/hosts associating your current hostname to 127.0.1.1. In case we change our domain, we need to modify this line. Example:

• my short host name is <vm_name>, and my new domain is .sa.cs.ut.ee, so the full domain name <vm_name>.sa.cs.ut.ee
• the line to be added should look something like this:
  • 127.0.1.1 <vm_name> <vm_name>.sa.cs.ut.ee

6. Configuring the BIND Name Server

First, install the Bind 9 DNS server-related packages.

Now let's create a directory for logging and necessary files within that directory that will be populated by the logging configuration you will add later in this lab.

As the BIND name server runs with the rights of the root user and group named (not root), you should always ensure that the main configuration file and zone files have the correct permissions and group ownership. Every file to be read by the BIND DNS server must be readable by the named group. If you get "permission denied " errors, use the following commands to set the owner or group ownership, for example:

• # chown root:named /etc/targetConfFile.conf

and following commands to set the correct permissions: (Learn More)

• # chmod 640 /etc/targetConfFile.conf
  • Will set rights rw- r-- ---
• # chmod g+r /etc/targetConfFile.conf
  • Will just set read right for the group
• Use chown to change owner/group. Use ls -la to check permissions/owner/group.

The main configuration file for BIND is :

• /etc/named.conf <-- Main configuration file - bind options, logging options, zones. This configuration could be split into smaller segments if necessary, but that means you would have to specify the include statement in the main configuration file.

Let's start configuring the DNS server by adding proper logging configuration. Replace the default logging options in the named.conf with the following:

logging {
     channel default_log {
          file "/var/named/log/default" versions 3 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity info;
     };
     channel auth_servers_log {
          file "/var/named/log/auth_servers" versions 100 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity info;
     };
     channel zone_transfers_log {
          file "/var/named/log/zone_transfers" versions 3 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity info;
     };
     channel client_security_log {
          file "/var/named/log/client_security" versions 3 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity info;
     };
//
// If you have the category 'queries' defined, and you don't want query logging
// by default, make sure you add option 'query log no;' - then you can toggle
// query logging on (and off again) using command' rndc querylog'
//
     channel queries_log {
          file "/var/named/log/queries" versions 600 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity info;
     };
//
// This channel is dynamic so that when the debug level is increased using
// rndc while the server is running, extra information will be logged about
// failing queries. Other debug information for other categories will be
// sent to the channel default_debug (which is also dynamic), but without
// affecting the regular logging.
//
     channel query-errors_log {
          file "/var/named/log/query-errors" versions 5 size 20m;
          print-time yes;
          print-category yes;
          print-severity yes;
          severity dynamic;
     };
//
// This is the default Syslog channel, defined here for clarity. You don't
// have to use it if you prefer to log to your own channels.
// It sends to syslog daemon facility and sends only logged messages
// of priority info and higher.
// (The options to print time, category, and severity are non-default.)
//
     channel default_syslog {
          print-time yes;
          print-category yes;
          print-severity yes;
          syslog daemon;
          severity info;
     };
//
// This is the default debug output channel, defined here for clarity. You
// might want to redefine the output destination if it doesn't fit with your
// local system administration plans for logging. It is also a special
// channel that only produces output if the debug level is non-zero.
//
     channel default_debug {
          print-time yes;
          print-category yes;
          print-severity yes;
          file "named.run";
          severity dynamic;
     };
//
// Log routine stuff to Syslog and default log:
//
     category default { default_syslog; default_debug; default_log; };
     category config { default_syslog; default_debug; default_log; };
     category dispatch { default_syslog; default_debug; default_log; };
     category network { default_syslog; default_debug; default_log; };
     category general { default_syslog; default_debug; default_log; };
//
// Log messages relating to what we got back from authoritative servers during
// recursion (if lame-servers and edns-disabled are obscuring other messages
// they can be sent to their own channel or to null). Sometimes these log
// messages will be useful to research why some domains don't resolve or
//don't resolve reliably
//
     category resolver { auth_servers_log; default_debug; };
     category cname { auth_servers_log; default_debug; };
     category delegation-only { auth_servers_log; default_debug; };
//
// Log together all messages relating to authoritative zone propagation
//
     category notify { zone_transfers_log; default_debug; };
     category xfer-in { zone_transfers_log; default_debug; };
     category xfer-out { zone_transfers_log; default_debug; };
//
// Log together all messages relating to client access and security.
// (There is an additional category 'unmatched' that is by default sent to
// null but which can be added here if you want more than the one-line
// summary that is logged for failures to match a view).
//
     category client{ client_security_log; default_debug; };
     category security { client_security_log; default_debug; };
//
// If you have the category 'queries' defined, and you don't want query logging
// by default, make sure you add option 'query log no;' - then you can toggle
// query logging on (and off again) using command' rndc querylog'
//
     category queries { queries_log; };
//
// This logging category will only emit messages at debug levels of 1 or
// higher - it can be useful to troubleshoot problems where queries are
// resulting in a SERVFAIL response.
//
     category query-errors {query-errors_log; };
};

Now we add the simple caching DNS configuration to our bind server. This is to get rid of forwarding the DNS resolve queries to 10.10.10.1. The caching DNS server running locally on the same host will do the trick. Instead of forwarding the queries to 10.10.10.1, we will use our own resolver on 127.0.0.1, which has enabled recursive resolving (and caching).

Edit the DNS configuration file named.conf:

listen-on port 53 { any; };
listen-on-v6 port 53 { none; };
...
allow-query     { any; };
...
recursion yes;
...
dnssec-enable no;
dnssec-validation no;

In the configuration, we declare that anyone should be allowed to send resolve queries to our bind server by having allow-query { any; }; defined as such. Secondly, we disable queries from IPv6, which we have disabled previously in section 5. In addition we increase the scope of IPv4 by putting bind listen on all interfaces' port 53 listen-on port 53 { any; }; Then we enable recursion and finally turn off dnssec. We turn off dnssec because enabling this would require admin-level access to ut.ee domain name servers, which we do not have.

The recursion is needed to let bind act as a standard caching recursive resolver (like those assigned by DHCP or put manually in /etc/resolv.conf).

• Example query "where is courses.cs.ut.ee" is then resolved as follows:
  • "where is .ee"
  • "where is ut.ee"
  • "where is cs.ut.ee"
  • "where is courses.cs.ut.ee"
• The often made queries are cached and answered explicitly

Now you can check your configuration using the named-checkconf utility

No output means configuration is OK.

Now query the service status and if needed, check the log files (those we have created in the beginning)

If everything works fine and no error is reported, you may check how your local resolver is performing:

If you have an error related to permissions to /var/named/log/default , please run the following command:

• restorecon -Rv /var/named
• Also make sure the file/folder permissions are correct.

This command relabels the security of that folder, allowing Bind to write there properly. Now systemctl start named should work properly.

It should answer with a list of IP addresses for courses.cs.ut.ee like this:

; <<>> DiG 9.11.4-P2-RedHat-9.11.4-26.P2.el8 <<>> @127.0.0.1 courses.cs.ut.ee
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 53403
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 2, ADDITIONAL: 4

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
; COOKIE: 4d3b69aaf5aa48ca539139635e63c827c5801f3d6572b073 (good)
;; QUESTION SECTION:
;courses.cs.ut.ee.      IN  A

;; ANSWER SECTION:
courses.cs.ut.ee.   4320    IN  CNAME   web.cs.ut.ee.
web.cs.ut.ee.       4320    IN  A   193.40.36.57
web.cs.ut.ee.       4320    IN  A   193.40.36.55

;; AUTHORITY SECTION:
cs.ut.ee.       4320    IN  NS  ns2.ut.ee.
cs.ut.ee.       4320    IN  NS  ns.ut.ee.

;; ADDITIONAL SECTION:
ns2.ut.ee.      10800   IN  A   193.40.5.76
ns.ut.ee.       10800   IN  A   193.40.5.99
ns2.ut.ee.      10800   IN  AAAA    2001:bb8:2002:500::76

;; Query time: 23 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Sat Mar 07 18:13:27 EET 2020
;; MSG SIZE  rcvd: 218

The local resolver is working as intended if you get an answer like this for courses.cs.ut.ee. You may want to enable it for auto-starting at the boot time:

The /etc/resolv.conf will not be rewritten by DHCP client (next time we reboot) and stays static. So we can safely edit it to rely on your own local resolver for doing DNS queries by replacing the namesever variable value with 127.0.0.1

Now your local resolver is used system-wide, so you can test it by just pinging some hosts by name:

Creating a Personal Domain for named

PS! Before continuing please make sure you have completed all of the previous steps.

Adding zone definition to the local resolver

Creating a new zone is a two-step process:

• First, the records for the zone must be entered into the zone file
• Second, the zone must be described/declared in the /etc/named.conf file.

Understanding the zone record files

The typical Zone file contains information about associations between hostnames and IP addresses within the domain of responsibility. The records inside the Zone file can be of different types:

• SOA - Start of Authority. Defines the zone name, email contact, and various time and refresh values applicable to the zone.
• A - IPv4 Address record. An IPv4 address for a host.
• AAAA - IPv6 Address record. An IPv6 address for a host. Current IETF recommendation for IPv6 forward-mapped zones.
• NS - Name Server. This record defines the authoritative name server(s) for the domain (defined by the SOA record) or the sub-domain.
• MX - Mail Exchange. A preference value and the hostname for a mail server/exchange that will service this zone. RFC 974 defines valid names.
• PTR - IP address (IPv4 or IPv6) to host. Used in reverse maps.
• CNAME - Canonical Name. An alias name for a host. Causes redirection for a single RR at the owner-name.

More details about record types can be found here.

Example Zone file content looks as follows:

NB! In Zone file ; denotes comments in configuration

;
; Zone file for domain-example.org
$ORIGIN domain-example.org.  ; Zone starting point in the name-space tree
                     ; (if skipped - the origin is inferred by the DNS server based on
                     ; zone declaration in name.conf.local file
$TTL 15M             ; Default expiration time for all resources
@ IN SOA ns.domain-example.org. root.domain-example.org.(
     2012020101       ; Serial
        15M       ; Refresh (15 minutes)
        5M        ; Retry (5 minutes)
        604800    ; Expire (7 days in seconds)
        600 )     ; Negative Cache TTL
@       IN      NS      ns            ; name server for domain-example.org
@       IN      A       172.17.6.1    ; IPv4 address for domain-example.org
ns      IN      A       172.17.6.1    ; IPv4 address for ns.domain-example.org
@       IN      MX  10  mail          ; mail server for domain-example.org
myhost  IN      A       172.17.6.5    ; IPv4 address for myhost.domain-example.org
mail    IN      A       172.17.6.6    ; IPv4 address for mail.domain-example.org
myhost2 IN      A       172.17.6.7    ; IPv4 address for myhost2.domain-example.org
www     IN      CNAME   myhost        ; www is an alias for myhost in domain-example.org
ftp     IN      CNAME   myhost        ; ftp is an alias for myhost in domain-example.org
www.other-domain.org. IN A 192.168.1.11 ;  www.other-domain.org. is outside of domain-example.org
                                      ; and therefore needs to be declared using FQDN

The syntax is explained here:

• ; - Comment, disabled line
• $ORIGIN - The zone's starting point in the namespace tree, (if skipped - the origin is inferred by the DNS server based on zone declaration in name.conf.local file
• $TTL - Default expiration time for all resources (cache time, specified in seconds)
  • (alternatively M, H might be appended to denote minutes or hours)
• @ IN SOA ns.domain-example.org. root.domain-example.org.( 2012020101 15M 5M 604800 600) :
  • @ - Will be expanded to $ORIGIN if specified or assigned by DNS server
    • (in current example to domain-example.org)
  • IN - Protocol family:
    • IN - Internet protocol (IP)
    • HS and CH - historic MIT protocols from the times there was DNS, but there was no IP protocol.
  • SOA - Zone authority info:
  • ns.domain-example.org - Address of name-server
  • root.domain-example.org - Will be expanded to root@domain-example.org and will correspond to the email address of DNS master in the domain-example.org
  • 2012020101 - Serial Number (as specified here):
    • Unsigned 32 bit value in range 1 to 4294967295 with a maximum increment of 2147483647. In BIND implementations, this is defined to be a ten-digit field. This value MUST increment when any resource record in the zone file is updated. A slave (Secondary) DNS server will read the master DNS SOA record periodically, either on expiry of refresh (defined below) or when it receives a NOTIFY and compares, arithmetically, its current value of sn with that received from the master. If the sn value from the master is arithmetically HIGHER than that currently stored by the slave, then a zone transfer (AXFR/IXFR) is initiated by the slave. If the value of sn from the master DNS SOA is the same or LOWER, then no zone transfer is initiated.
    • The convention is to use a date-based sn value to simplify the task of incrementing the sn - the most popular convention being yyyymmddss where yyyy = year, mm = month, dd = day and ss = a sequence number in case you update it more than once in the day.

As you can see after declaring SOA Zone authority, the actual Zone related data is added: the zone records:

• <name> <ttl> <class> <type> <data>
  • name record name
  • ttl time to live in seconds, how long cache lives, if skipped $TTL variable is used.
  • class record class, the protocol family (in modern world only IN remains)
  • type record type
  • data type related value of the record

Record name can be one of three:

• Short hostname (inside the current domain), like ns or mail , which will be then automatically extended to ns.domain-example.org and mail.domain-example.org.
• FQDN (Fully-Qualified Domain Name), like www.otherd-domain.org.
  • "'In this case, we have to terminate FQDN with . to avoid auto-expansion to www.other-domain.org.domain-example.org '''
• @ sign which will be expanded to domain-example.org denoting the domain itself.

Create your zone file

You need to create the zone file in /etc/named/, but first, let's make sure you know your domain:

Here we assume you got your "domain name" correctly, and we proceed with Zone file creation:

If the file was created properly, you may edit the zone and add the bare minimum amount of zone records to your zone file.

Example file of ours /etc/named/student-test.sa.cs.ut.ee , domain student-test.sa.cs.ut.ee and IP 172.17.64.a :

;
; BIND data file for local zone student-test.sa.cs.ut.ee;
$TTL    15M
@       IN      SOA     ns1.student-test.sa.cs.ut.ee. root.student-test.sa.cs.ut.ee. (
                     2012020201         ; Serial
                            15M         ; Refresh
                             5M         ; Retry
                           120M         ; Expire
                            600 )       ; Negative Cache TTL
@              IN      NS      ns1
@              IN      A       172.17.64.a
ns1            IN      A       172.17.64.a
student-test   IN      A       172.17.64.a

If the zone file is correct, we may add the defined zone into the Bind configuration.

In our examples, we made it as follows:

...
zone "student-test.sa.cs.ut.ee" IN {
        type master;
        file "/etc/named/student-test.sa.cs.ut.ee";
        allow-update { none; };
};

If the configuration is OK:

The log has to have a line like:

...
08-Mar-2020 16:35:58.841 general: info: zone student-test.sa.cs.ut.ee/IN: loaded serial 2020030804
08-Mar-2020 16:35:58.841 general: notice: all zones loaded
08-Mar-2020 16:35:58.841 general: notice: running...
...

Afterwards you should be able to use dig or nslookup commands for testing:

Some of the queries might return ANSWER: 0? Can you explain that?

Rerun the queries:

Example dig result of the command:

• dig @127.0.0.1 student-test.sa.cs.ut.ee

; <<>> DiG 9.11.4-P2-RedHat-9.11.4-26.P2.el8 <<>> @127.0.0.1 student-test.sa.cs.ut.ee
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 64197
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 2

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
; COOKIE: 5b6c2462c45d2d8c86684aec5e65045a42fab1a5ec27af6a (good)
;; QUESTION SECTION:
;student-test.sa.cs.ut.ee.      IN  A

;; ANSWER SECTION:
student-test.sa.cs.ut.ee.   900 IN  A   172.17.65.65

;; AUTHORITY SECTION:
student-test.sa.cs.ut.ee.   900 IN  NS  ns1.student-test.sa.cs.ut.ee.

;; ADDITIONAL SECTION:
ns1.student-test.sa.cs.ut.ee.   900 IN  A   172.17.65.65

;; Query time: 2 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Sun Mar 08 16:42:34 EET 2020
;; MSG SIZE  rcvd: 125

What notify explicit; does, is that it forces the DNS server to notify slave DNS servers about updates. Without it, the other DNS servers might not pick up updates to your domain before the end of TTL time, which can sometimes take very long.

"'Understanding reverse mapping Zone files'"

The interesting fact about records in Zone file configurations:

• all the associations are unidirectional

which means if we have

• ns1.example.org. A IN 10.10.10.5

it is equivalent to the link

• ns1.example.org. -> 10.10.10.5

and not opposite way around:

• 10.10.10.5 -> ns1.example.org.

The secondary link from IP to the hostname has to be added by introducing the PTR records. The PTR records are organized and collected in a separate Zone file called Reverse Mapped Zone.

You already know the DNS namespace is organized in a hierarchical structure:

The whole domain name space is first addressed from root(.) which is the root of the whole DNS tree, then it's divided across number of TLDs, like .com, .org, .net etc. You can check the full list of TLDs here. Next, each TLD has several sub-domains, and for us, we did introduce additional TLD <name>.sa.cs.ut.ee with sub-domains for each student. Following the "divide and conquer" idea, each of the sub-domains can further be divided into sub-domains (each student can potentially create an infinite amount of sub-domains).

Considering the TLDs, there is one specific:"' .arpa'" which currently isn't serving any of the domains but is holding routing and addressing parameters and is therefore called Infrastructure TLD.

• The DNS architecture was designed by US Advanced Research Project Agency (ARPA)
• "' .arpa'" was originally the domain name of the ARPA agency, and it was the first domain to be introduced in the DNS namespace.
• as it turned to be Infra-TLD, the "'.arpa'" is referred to as Addressing and Routing Parameter Area.

The idea behind "'.arpa'" Infra-TLD is to provide reverse mappings (from IP to hostname) and is mainly used by humans for such things as tracking where a website visitor came from or where an email message originated etc.

However, reverse DNS is important for one particular application. Many email servers on the Internet are configured to reject incoming emails from any IP address which does not have reverse DNS. So if you run your own email server, reverse DNS must exist for the IP address that the outgoing email is sent from. Reverse DNS mappings are also an important part of keeping the trust-based DNS working. A simple check any service provider can run is looking for mismatches in PTR and A type records, and usually, if these don't match, data is marked as spam.

The IPv4 address space is, of course, not infinite, but the idea of hierarchy is followable if we consider how the IPv4 space is divided into subnets. Each organization is assigned with a big subnet (A class or B class dependent on the size of the organization). Dependent on the inner structure of the organization, the big subnet can be divided again and again till there are free unused IPv4 addresses.

As you can see, the bigger subnets can still be broken into smaller ones; the simple example is explained here (if you struggle understanding concepts of IP addressing, you may first read the IP Basics). Now, taking this into account, all of the IP address space was structured and reduced into reverse mapping Zones under the"' .arpa'" TLD. Therefore the .arpa is broken into sub-domains as follows:

• in-addr.ARPA domain contains all the reverse mappings for IPv4 address space ("in-addr" stands for Internet Address)
• in-ip6.ARPA domain contains all the reverse mappings for IPv6 address space

Now each of them is broken down into sub-domains denoting subnets of IPv4 or IPv6. The domain name for the subnet is then constructed as follows:

• For the IPv4 subnet:
  • Consider your subnet (for example, you have 172.17.6.0/24)
  • The non-changeable part is 172.17.6. (by subnet 172.17.6.0) and only the last octet can change (by netmask 255.255.255.0)
  • Non-changeable part is then reversed (so we get 6.17.172)
  • in-addr.arpa is appended to the reversed subnet (we get 6.17.172.in-addr.arpa)
  • 6.17.172.in-addr.arpa is then our domain for reverse mapping:
    • This will be the $ORIGIN for the reverse mapping Zone file

Now the reverse mapping Zone files are similar to regular Zone files. Here is an example:

$ORIGIN 6.17.172.in-addr.arpa.   ; Zone starting point in the name-space tree
                     ; (if skipped - the origin is inferred by the DNS server based on
                     ; zone declaration in name.conf.local file
$TTL 15M             ; Default expiration time for all resources
@ IN SOA ns.domain-example.org. root.domain-example.org.(
     2012020101       ; Serial
        15M       ; Refresh (15 minutes)
        5M        ; Retry (5 minutes)
        604800    ; Expire (7 days in seconds)
        600 )     ; Negative Cache TTL
@              IN      NS      ns.domain-example.org.       ; Name Server for domain-example.org
1             IN      PTR     ns.domain-example.org.       ; 172.17.6.1 has name ns.domain-example.org
5             IN      PTR     myhost.domain-example.org.   ; 172.17.6.5 has name myhost.domain-example.org
6             IN      PTR     mail.domain-example.org.     ; 172.17.6.6 has name mail.domain-example.org
7             IN      PTR     myhost2.domain-example.org.  ; 172.17.6.7 has name myhost2.domain-example.org

As you can see, there is not much change in the headers:

• $ORIGIN - pointing to 17.172.in-addr.arpa
  • we can omit specifying origin, leaving it for bind9 to specify it (inferring named.conf.local configurations
• TTL and SOA records remain the same as in the regular Zone file (we are still in domain-example.org)
• NS record is the same (we are still having DNS server's FDQN: ns.domain-example.org.)
• Next, we have the PTR record for each A or AAAA type of record from the regular Zone file of our domain
  • Each hostname -> IP association should have a reverse association IP -> hostname.

Please note that it is possible to specify multiple A or AAAA types of records pointing to the same IP address, however in reverse mapping, only one of the hostnames may be pointed by the PTR type of record. For example, in the regular Zone file, we have:

@       IN      A       172.17.6.1    ; IPv4 address for domain-example.org
ns      IN      A       172.17.6.1    ; IPv4 address for ns.domain-example.org

Both "domain-example.org" and "ns.domain-example.org" have the A type of record pointing to the same IP of 172.17.6.1. In the reverse mappings, however, we have only one record:

1             IN      PTR     ns.domain-example.org.

... saying 172.17.6.1 has name ns.domain-example.org Please keep in mind this is an example. For you, it needs to be a different IP address.

The content of the reverse mapping Zone file is then stored similar to the regular zone:

• Zone example-domain.org was stored in file example-domain.org.zone
• Zone 6.17.172.in-addr.arpa should be stored in file 6.17.172.in-addr.arpa.zone

The filename convention is not important to follow (you can use other names if you wish). The idea of this convention is to keep named.conf.local configuration file more clear and followable.

The last thing to understand is how the reverse mapping Zone file is embedded into the existing bind9 configuration, and this is fact 1-to-1 similar to how we declared the regular zone in bind9 configuration (we have to add another zone clause in the "configuration" file):

Declaring the zone clause for reverse mapping Zone is then looking like:

zone "65.17.172.in-addr.arpa" IN {
        type master;
        file "/etc/named/reverse.student-test.sa.cs.ut.ee";
        allow-update { none; };
};

After the reverse mapping Zone file is included in bind configuration, the standard check-routines should follow:

• Checking the reverse mapping Zone file:
  • named-checkzone ''65.17.172.in-addr.arpa'' <path to reverse.student-test.sa.cs.ut.ee file>
• Checking the named configuration using named-checkconf command

Now we are ready to add the reverse mapping Zone to our bind9 configuration for our network and to populate it with PTR records.

If the reverse mapping Zone is OK, construct the reverse mapping domain name:

If you have calculated the reverse mapping name for your subnet:

If the configuration is OK:

7. Ansible - automating the above

The focal point of the DNS lab for Ansible is mostly getting comfortable with template module. There is also the copy module. But due to the usability aspect of templates, we suggest using the template module over the copy module.

You are not obliged to use templating if the copy module suits you more but read this artcile from Anthony Critelli at redhat.com to give you a proper distinction between the two.

"'TL;DR version: Copy module copies files from the local host to a remote host that you want to manage. Template module also copies files from the local host to remote host but allows you to dynamically change the template's content through variables, Jinja2 is the magic behind the template module.'"

Let's start by making sure the structure is there:

The structure should look something like this:

|
|- playbook.yml
|- roles/
|    |- dns
|    |   |- templates
|    |   |- tasks
|    |        |- main.yml
...

You should now start populating the roles/DNS/tasks/main.yml file. The following example is written for you based on the above lab manual. The general idea is that once you know how to manually disable IPv6 in the lab VM, you should also write it into your playbook. The following example has been purposefully left halfway done; it is here to get you started.

---
- name: DNS | Autoremove unneeded packages installed as dependencies
  dnf:
    autoremove: yes

- name: Disable IPv6
  copy:
    src: 70-ipv6.conf
    dest: /etc/sysctl.d/70-ipv6.conf
    owner: root
    group: root
    mode: '0644'

- name: Change net.ipv6.conf.default.disable_ipv6 to 1.
  sysctl:
    name: net.ipv6.conf.default.disable_ipv6
    value: '1'
    sysctl_file: /etc/sysctl.d/70-ipv6.conf
    reload: yes

- name: Change net.ipv6.conf.all.disable_ipv6 to 1.
  sysctl:
    name: net.ipv6.conf.all.disable_ipv6
    value: '1'
    sysctl_file: /etc/sysctl.d/70-ipv6.conf
    reload: yes

- name: Inserts into hostname
  template:
    src: hostname.j2
    dest: /etc/hostname
    owner: root
    group: root
    mode: '0644'

- name: Inserts into hosts
  template:
    src: hosts.j2
    dest: /etc/hosts
    owner: root
    group: root
    mode: '0644'

If you run this, Ansible will fail when it gets to processing the template modules. This is because we have not made the template files that it needs to use to create the files on the server.

The template module searches files from the template/ folder, and similarly, the copy module searches files from the files folder.

If you are running the modules in a role, then first the role's template and files folders are checked. If nothing is found there, then the whole Ansible repository's files and folders are checked. If nothing is found there either, then the folders in /etc/ansible/templates and /etc/ansible/files are searched. This is called precedence and is important for grouping information. You would want to have files related to the DNS service in the DNS role's directory, while files specific for a whole group/environment would probably make sense to be in the Ansible repository's files/folders.

|- templates
|- playbook.yml
|- roles/
|    |- dns
|    |   |- templates
|    |   |    |- hostname.j2
|    |   |    |- hosts.j2
|    |   |
|    |   |- files
|    |   |    |- 70-ipv6.conf
|    |   |       
|    |   |- tasks
|    |        |- main.yml
|    |   |- files
...

Let's add the templates files, but by utilizing some of the intelligence Ansible allows us to apply. In the following part, you will notice we are using variables. Having {{ hostname }} and {{ domain_name }} as variables is useful for the lifecycle of this playbook, as during the exam, these variables could change, and then instead of having to change all the files, you can only edit one place.

net.ipv6.conf.default.disable_ipv6=1
net.ipv6.conf.all.disable_ipv6=1

127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
127.0.1.1   {{ hostname }}.{{ domain_name }} {{ hostname }}
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6

{{ hostname }}.{{ domain_name }}

The only thing left to do is to also add the variable definitions.

---
- hosts: localhost
  user: centos
  become: yes
  vars:
    hostname: B123123
    domain_name: sa.cs.ut.ee

  roles:
   - { role: etais, tags: etais } 
   - { role: dns, tags: dns }

The role we wrote so far does not cover the whole lab. The point of this section was to get you introduced to all the tools. From this point onward, you should use the template or copy modules for these files: <your_domain>, reverse.<your_domain> , resolv.conf, named.conf. You should use the systemd module to restart named.

Another thing to take into consideration with our virtual machines is that you cannot see the machine's public IP address (172.17.XX.XXX) inside the machine. This is because of how the cloud infrastructure is set up.

You can solve this by either using a self-defined variable for the IP address or by doing some bit more complicated logic. OpenStack has something called a metadata server, that answers with information about the host. We can use this inside the Ansible to receive the IP address.

  - name: Get OpenStack metadata
    uri:
      url: http://169.254.169.254/latest/meta-data/public-ipv4 
      method: GET
      status_code: 200
      return_content: yes
    register: cloud_metadata

  - name: Print IP address
    debug: var=cloud_metadata.content

8. Keeping your Ansible repository safe in our Gitlab

And as always, push your stuff to our Gitlab.