New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
dnsmasq: Maximum number of concurrent DNS queries reached (max: 150) #166
Comments
you can raise default (150) value of concurrent queries:
|
you can also use consul as your default DNS instead of dnsmasq by:
|
Ok, thanks @sielaq! I will try to update that change. The weird thing is, I am not sure why I am suddenly getting this error. I haven't added any new servers to the cluster, so I wouldn't expect any more DNS traffic than usual. Is there anything inside of PanteraS that would be flooding dnsmasq? Lastly, do you know if there's a way to look closer at the dnsmasq logs to see which host is causing the problem? |
if this is caused by one of your service then yes, it looks like your service is flooding. |
can you trace what exactly is flooding ? |
Strange because I don't actually have any services running and I still get the error. |
Can I look at the dnsmasq logs to determine which server is the culprit? I've only got 2 slaves and 1 master and they're all configured to use the master as the DNS server... so I wouldn't expect it to be flooded... |
|
yea you can log queries too, additional |
Thanks! I ran the commands and originally saw a bunch of traffic (probably outbound) going to my corporate DNS servers. So I removed the corp DNS servers from I never saw any traffic (using either of the commands), yet, I'm still experiencing the error. I have stopped and removed the PanteraS container, but I'm still seeing the issue. |
I'm actually seeing the error immediately after startup:
How can dnsmasq be saturated after it has only been running for less than 5 seconds? |
" I have stopped and removed the PanteraS container, but I'm still seeing the issue." |
Yes, I believe so. Right now I've only got a single server (master) running. When I start fresh with a new |
login into the PaaS container like if all is stable, stop everything except |
Can you confirm the following requests are showing inbound to
My master node is |
This is rather the response, 53 answer to port 19762. |
ahh 10.7.112.99 is your PanetraS. Did you try to stop mesos/marathon/zookeper with supervisorctl as I asked ? |
Yes,
I stopped each of the services, one by one, until only dnsmasq was left running:
At this point, I was still receiving requests on port 53 as shown in my previous comment. So I stopped dnsmasq, and then the requests stopped. When I started dnsmasq again, I did not see the requests coming in. So I started each of the services one by one, and started noticing the flood immediately after starting |
I stopped just consul and dnsmasq and the flooding stopped. I then started dnsmasq and didn't see the flooding. Only after starting consul do I see the flooding... |
I went back to a build before the consul upgrade to |
I reverted to commit eec2c7673ebd49aa6a240234725d07aa79328e3d (before consul was updated to 0.6), and everything works as expected. Are you experiencing this issue? |
Ok, after hours of switching between commits, I finally got it working on the latest build. I haven't changed anything, so I'm not sure what has happened. I opened up another issue asking about proper DNS configuration. I'll close this one for now. Thanks for the help @sielaq! |
Ok, actually, I'm reopening this issue because I've gotten to the point where I can get my masters and slaves up and running without the DNS flood. However, the first docker container I spawn with Marathon causes the issue. I start a docker image with marathon and it spawns a container ID of
As you can see, the query in the above snippet is referencing Even after I stop and remove the container, I still continue to see the DNS floods with the same reference to Do you have any ideas of why this would be happening? |
Yea I also don't think it is a consul issue. my 2 hypotheses :
|
I tend to agree with your theory about the DNS loop. Is there any way to tell my PanteraS hosts to only send requests to the corporate DNS servers if they are not service.consul lookups? Here is the output of my master (IP address 10.0.0.10)
slave 1 (IP address 10.0.0.11)
slave 2 (IP address 10.0.0.12)
You are correct. Removing the |
Actually, I retract my last comment. Removing the |
More specifically, it seems to be the |
I was able to reproduce it. |
Actually, it looks like The only change I have made is the Have you noticed the same? Should I open a new issue? |
… interfaces generate loop
I will fix it with this issue. |
Hi,
For some reason, recently, I've started experiencing the following error:
As a result, my lookups are failing for my marathon services. Do you have any idea what could be causing this, or if there's a way to look at the dnsmasq logs to determine where the queries are coming from?
Thanks!
The text was updated successfully, but these errors were encountered: