Replication issues #105
Comments
|
Try to see if there's an error in the master or slave logs. Do you have set supermaster correctly? https://github.com/pschiffe/docker-pdns/blob/master/docker-compose.yml#L88 |
|
I respected what you did! I use swarm cluster : |
|
Thank you; could you also share logs from master and slave containers, from the start of the containers to after you create the zone? If there's an error in communication, it should be visible there. |
|
Thank you for your help : |
|
I think this is the mistake |
|
After you create a new zone, you also need to add Once you add the |
|
It's all the same, I have done as you said but my zone has not spread to the slaves and there are no new logs yet.
|
|
|
|
That's weird, I'm not sure here. Try to create a new random record in the |
|
|
Can you verify connectivity between the containers? ie exec to ns1, and try Also you can try to provide the master env var |
|
I added this line |
|
The stdout output of the |
|
|
|
You were right, it is the network that is the problem. |
I actually notice that my services are not getting the ip I assigned in the YAML, there is |
|
Now I see that it's not possible to use static IPs for containers in docker swarm: |
|
At least supermaster feature requires static IP, and axfr transfers probably as well; I don't see right now how this could be work-arounded for swarm :/ |
|
How are you? Thanks for your help, I've been looking for how to define static ip's to services with swarm, but I couldn't find anything. I tried to modify I can do curl inside the containers. My problem is that I have to fill in the fields Is it possible to proceed in this way: All services start except pdns-admin-static |
|
hi @asow25, I'm good, thank you. The error from static container means that Reg. the options, the Problem is with the SUPERMASTER IPS, there must be an IP: https://doc.powerdns.com/authoritative/modes-of-operation.html#autoprimary-automatic-provisioning-of-secondaries The slave replication can work without SUPERMASTER feature, but you have to provision the domain on slaves manually (with NS records). The third option can be to don't use pdns replication, but use mysql replication instead. That's a supported mode of operation: https://doc.powerdns.com/authoritative/modes-of-operation.html#native-replication Otherwise I'm not sure, as the static IPs are not supported for docker swarm: moby/moby#24170 and we are dealing with DNS here 😅 |
|
I managed to set static ip to and a second network for the rest of the services At first, I had errors that it was the same gateway so I changed it to 192.168.100.2 When I run the stack, it's Ok but when I connect to this means my two networks are not talking to each other but overlay is the swarm network that allows multi-data communication. Anyway, I added an alias I still have the same error, my two networks do not communicate so the databases of these three services were not created when I deploy the stack. Can you help me please ? |
|
I think 2 networks should be fine, but use different subnets, as this will most likely clash. You can attach container to multiple networks. Also alias must be different for every container. |
|
It's all the same, Please take a look! |
|
hi @asow25, I'm sorry I cannot tell what could be the problem here. You can try to exec to the containers and try some network debugging to check if the networks are routing correctly. |
|
Hi @pschiffe all my services are up and running, my 3 powerdns have static ip's and a second ip that allows them to communicate with other services. I checked the container logs, everything seems to be working fine except When I try to create a zone, I get an error : |
|
If the pdns-master is not available under |
|
|
The value must include full URL, with |
|
It's the same, honestly, I don't understand, when I fix a problem, another one is created. Before everything worked, but the ip were not fixed so impossible that a zone is propagated to other servers powernds since swarm manages itself the allocation of ip addresses. I found this method: Now the ip's are fixed and I have added another common network for all the services to communicate. Then, database problem, it's fixed, now it's the turn of |
|
I'm so sorry to hear that :( It's true that docker swarm doesn't have very good reputation, maybe you could take a look at some simple kubernetes distro, such as https://k3s.io/ |
|
hi @asow25, updated images have been pushed to the registry, please update and try again. the issues with quotes should be hopefully resolved now |
|
Many thanks! One question, which syntax should I use, |
|
Thanks a lot, it works, I have two zones on pdns-admin. The zones are on my galera nodes |
|
I have a question. I have a Now that I have only one |
|
Awesome, that's great to hear @asow25!
It doesn't matter anymore, single quotes are fine now. Numbers can also be without quotes.
With Native, replication is handled with your galera cluster, so powerdns doesn't care anymore. Now, the |
|
Yes I am talking about client side, I had 02 theories:
What is the procedure to follow ? |
|
If you have only one instance of |
|
I noticed that pdns-admin after several minutes, I have this error: |
|
Pls post the logs of uwsgi and nginx containers after you trigger the error. |
|
|
When my services were all in replicas 1, there was no problem. I made a scale of the services pdns-master and pdns-recursor replicas=3 |
|
It's all about MySQL connections - |
|
I removed galera and dldb, I bump into https://github.com/colinmollenhour/mariadb-galera-swarm |
|
|
|
It seems like mysql connections are idling. You can try to increase the |
|
I have executed the commands |
|
I think that means that |
|
I found this |
|
For the mariadb server? Feel free to try it. |
|
I looked at the official site of mariadb, I read the doc about wait_timeout The value has changed, even when I log out and log in again, I have launched pdns-admin, I am waiting to see |
|
Thank you so much, I can't thank you enough, you don't know anything about me and yet you give me your time, you help me a lot and that's very nice! |
|
you are welcome ❤️ |
|
I have the same mistakes |
|
You can try to take a look in the pdns admin project, or ask there: https://github.com/search?q=repo%3APowerDNS-Admin%2FPowerDNS-Admin+Lost+connection+to+server+during+query&type=issues |
|
Thank you, I will try |
|
Hi @pschiffe I have attempted to set up pdns on k8s, but I'm encountering issues with the configuration. When I comment out the line Here's my manifest file |
|
@asow25, is this also resolved? |
|
@pschiffe, yes thank you! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I've got the issues with replication
When I create a zone on the master, the zones do not replicate on the slave. With php admin, I see my zones in powerdns, powernds-admin but in powerdns-slave > Domains, there is nothing. Is this normal? I would like it to replicate?
The text was updated successfully, but these errors were encountered: