Use ACK CoreDNS DNSTAP Analyser to diagnose DNS resolution errors - Container Service for Kubernetes

Container Service for Kubernetes (ACK) provides the ACK CoreDNS DNSTAP Analyser component that can be used to identify and classify DNS messages that are exchanged between a client and CoreDNS, or between CoreDNS and an upstream DNS server. You can use this component to identify the causes of DNS resolution errors in an efficient manner.

Prerequisites

An ACK Pro cluster is created. For more information, see Create an ACK Pro cluster.
The kubeconfig file of the ACK Pro cluster is obtained and the cluster is connected by using kubectl. For more information, see Obtain the kubeconfig file of a cluster and use kubectl to connect to the cluster.

Background Information

dnstap is a log format for DNS messages. You can use dnstap to diagnose DNS resolution errors on DNS servers. For more information, see dnstap.

Step 1: Install ACK CoreDNS DNSTAP Analyser

Log on to the ACK console. In the left-side navigation pane, click Cluster.
On the Clusters page, find the cluster that you want to manage and click its name. In the left-side pane of the cluster details page, choose Applications > Helm.

On the Helm page, click Deploy. In the Basic Information step, configure the parameters based on the following table.

Parameter	Example
Application Name	ack-coredns-dnstap-analyser
Namespace	kube-system
Source	Default value: Marketplace.
Chart	Select All for the Use Scenarios parameter. Select amd64 for the Supported Architecture parameter. Enter ack-coredns-dnstap-analyser in the search box.

Select ack-coredns-dnstap-analyser and click Next.

In the Parameters step, configure the Chart Version parameter and click OK.

Step 2: Configure the DNSTAP plug-in in CoreDNS

CoreDNS is pre-installed with the DNSTAP plug-in to send DNSTAP messages. The DNS messages that are sent and received by the DNSTAP plug-in can be forwarded to ACK CoreDNS DNSTAP Analyser. Before you can use ACK CoreDNS DNSTAP Analyser to diagnose CoreDNS resolution errors, you must configure and enable the DNSTAP plug-in in CoreDNS. After the DNSTAP plug-in is configured, CoreDNS forwards the DNS messages that are sent and received by the DNSTAP plug-in to ACK CoreDNS DNSTAP Analyser for automated diagnostics.

Run the following command to query and record the cluster IP address of ACK CoreDNS DNSTAP Analyser in the kube-system namespace. This IP address is used to configure the DNSTAP plug-in. In this example, the IP address is 172.21.0.10.
```
kubectl -n kube-system get svc dnstap-analyser
```
Run the following command to modify the ConfigMap of CoreDNS:
```
kubectl -n kube-system edit cm coredns -o yaml
```

Modify the default settings of CoreDNS based on the comments in the following snippet.

In the ConfigMap, you must enable the DNSTAP plug-in and set its destination address to the cluster IP address of ACK CoreDNS DNSTAP Analyser. This way, the DNS messages that are sent and received by the DNSTAP plug-in can be forwarded to ACK CoreDNS DNSTAP Analyser.

Corefile: |
        .:53 {
            #Details are omitted.
            ready

            # Add the code. Replace 172.21.0.10 with the actual cluster IP address of CoreDNS DNSTAP Analyser. 
            dnstap tcp://172.21.0.10:6000 full

            kubernetes cluster.local in-addr.arpa ip6.arpa {
            #Details are omitted.
            }
            #Details are omitted.
            reload
            loadbalance
        }

Run the following command to query all pods that run CoreDNS:

kubectl -n kube-system get pod | grep coredns

Expected output:

coredns-7d56l         1/1     Running   0          30m
coredns-s7m2t         1/1     Running   0          30m

Run the following command to query the CoreDNS log:
```
kubectl -n kube-system logs -f --tail=500 coredns-7d56l
```
Note
Replace coredns-7d56l with the name of a pod that is returned in Substep 4 of this step.
Wait for 1 minute. If the following output is returned and no exceptions occur, the configuration is modified.
```
[INFO] Reloading complete
```

Step 3: View the log of ACK CoreDNS DNSTAP Analyser

Run the following command to query all pods that run ACK CoreDNS DNSTAP Analyser:

kubectl -n kube-system get pod | grep dnstap-analyser

Expected output:

dnstap-analyser-bbdf879-g****         1/1     Running   0          30m

Run the following command to show the log of ACK CoreDNS DNSTAP Analyser:
```
kubectl -n kube-system logs -f dnstap-analyser-bbdf879-g****
```
Note
Replace dnstap-analyser-bbdf879-gpfkm with the name of a pod that is returned in Substep 1 of this step.

Step 4: Analyze the log of ACK CoreDNS DNSTAP Analyser

In the log of ACK CoreDNS DNSTAP Analyser, each line includes the aggregated information about a DNS session. A DNS session contains all requests and responses that are exchanged among a client, CoreDNS, and an upstream DNS server. A DNS session includes the following fields:

Status: The value of the Status field varies based on the scenario.

Status	Description
Succeeded	The domain name was resolved. The response carries the NXDOMAIN or NOERROR status code.
Failed	The domain name failed to be resolved. The response carries an error status code, or the connection to the upstream DNS server timed out.
SampleLoss	CoreDNS or ACK CoreDNS DNSTAP Analyser is overloaded and drops DNSTAP messages. In this case, you must increase the number of pods that run CoreDNS or ACK CoreDNS DNSTAP Analyser to reduce the load on each pod. This improves the accuracy of diagnostic results.

BitMap: indicates the type of the DNSTAP message that is captured from the session. You can identify possible causes of DNS resolution errors based on message types.

BitMap	Description
1	CoreDNS received a DNS query from the client but did not return a response to the client.
3	The internal domain name was resolved. The DNS query sent from a client hit the local DNS cache of CoreDNS or a DNS record in the Kubernetes cluster. The result was returned to the client.
5	The upstream DNS server failed to be connected. CoreDNS forwarded the DNS query sent from a client to the upstream DNS server. The upstream DNS server did not return a result.
15	The external domain name was resolved. CoreDNS forwarded the DNS query sent from a client to the upstream DNS server. The upstream DNS server resolved the domain name and returned the result to CoreDNS. Then, CoreDNS returned the result to the client.

Messages: This field records the content of the original DNS query and response in JSON format. You can analyze the message content based on dns-parameters-6.

(Optional) Step 5: Uninstall ACK CoreDNS DNSTAP Analyser

After you identify the cause, we recommend that you uninstall ACK CoreDNS DNSTAP Analyser to avoid unnecessary costs. To uninstall ACK CoreDNS DNSTAP Analyser, perform the following steps:

Delete the line of code that contains dnstap from the ConfigMap of CoreDNS:

Run the following command to modify the ConfigMap of CoreDNS:
```
kubectl -n kube-system edit cm coredns -o yaml
```

Delete the line of code that contains dnstap from the ConfigMap of CoreDNS based on the following content:

Corefile: |
        .:53 {
            #Details are omitted.
            ready

            # Delete the following line that you added.
            dnstap tcp://10.10.10.10:6000 full

            kubernetes cluster.local in-addr.arpa ip6.arpa {
            #Details are omitted.
            }
            #Details are omitted.
            reload
            loadbalance
        }

Run the following command to query all pods that run CoreDNS:

kubectl -n kube-system get pod | grep coredns

Expected output:

coredns-7d56l         1/1     Running   0          30m
coredns-s7m2t         1/1     Running   0          30m

Run the following command to query the CoreDNS log:
```
kubectl -n kube-system logs -f --tail=500 coredns-7d56l
```
Note
Replace coredns-7d56l with the name of a pod that is returned in preceding Substep c.
Wait for 1 minute. If the following output is returned and no exceptions occur, the configuration is modified.
```
[INFO] Reloading complete
```

Uninstall ACK CoreDNS DNSTAP Analyser.
1. Log on to the ACK console. In the left-side navigation pane, click Cluster.
2. On the Clusters page, find the cluster that you want to manage and click its name. In the left-side pane, choose Applications > Helm.
3. In the left-side pane of the cluster details page, choose Applications > Helm.
4. On the Helm page, find ack-coredns-dnstap-analyser and click Delete in the Actions column. Follow the on-screen instructions to delete ack-coredns-dnstap-analyser: