Massive spam of DNS requests for my local domains
Suddenly m network stopped working and by analyzing the symptoms i quickly figuredout it was related to the DNS.
Analyzing my router i found out a spam of DNS requests coming out from Homarr, pointing to my local domains (of my applications)
Solution:Jump to solution
Ok, then let's just keep tracking this in the GitHub issue as we discussed.
85 Replies
Thank you for submitting a support request.
Depending on the volume of requests, our team should get in contact with you shortly.
⚠️ Please include the following details in your post or we may reject your request without further comment: - Log (See https://homarr.dev/docs/community/faq#how-do-i-open-the-console--log) - Operating system (Unraid, TrueNAS, Ubuntu, ...) - Exact Homarr version (eg. 0.15.0, not latest) - Configuration (eg. docker-compose, screenshot or similar. Use ``your-text`` to format) - Other relevant information (eg. your devices, your browser, ...)
Frequently Asked Questions | Homarr documentation
Can I install Homarr on a Raspberry Pi?
docker logs for homarr
As requested in #🦞・general , please post logs, screenshot of the DNS TTL and other relevant info
Also, are you using a public DNS or a local DNS?
is the file fine or you want pastebin?
public, cloudflare 1.1.1.1
Pastebin, files cannot be opened on mobile and are suspectable to malicious code execution
Logs posted.
OS: proxmox debian 12> debian 12 VM> OMV> docker>homarr.
Homarr version 1.20.0
DNS request and reply captured with wireshark have a TTL of 64 - 63
Can you measure how many DNS packets you see per minute? It is normal that Homarr will perform multiple HTTP requests per second but DNS will be cached if TTL is high enough.
i do not have a counter, i am using my eye and i can see homarr makes a request for each subdomain i have, about one request every 2 seconds
reading the time too, ofc
One request per subdomain per 2 seconds? Or one request per two seconds but different subdomain?
Have you tried increasing TTL to a higher value?
Unless you have a specific use case, it's also not worth it to choose low TTLs:
https://blog.apnic.net/2019/11/12/stop-using-ridiculously-low-dns-ttls/
APNIC Blog
Stop using ridiculously low DNS TTLs | APNIC Blog
Guest Post: Why does half the Internet use a TTL of 1 minute or less?
one request per subdomain every 2 seconds-ish
so i have got like 10 apps, so 10 requests every 2 seconds
When checking the DNS packets, did you check those which were sent by Homarr or by your client?
i sniff the packets directly from the router, i see the packets sent and received from the IP of the VM and filtered by DNS
How do you run Homarr inside the VM? Any DNS client / cache in the VM?
OMV>docker managed with portainer
no DNS client, no custom cache
just basic network configuration with static ip and cloudflare dns
And just to confirm; you're not mixing up the TCP requests for the HTTP call with the DNS packets?
i have wireshark filtered for DNS, i dont see TCP and HTTP packets
https://i.imgur.com/md1PjZx.png an example
i was just forced to kill homarr now
Thanks for confirming. I can't observe the same behaviour on any of my setups nor my local development machine. Can you post on the GitHub issue https://github.com/homarr-labs/homarr/issues/1141 and post all details and reproduction steps? I suspect that you have a different network config causing caching to not work.
GitHub
bug: Request are not cached resulting in high network usage · Issu...
Provide environment information any Describe the bug Some queries are not cached but should. This results in an insane amount of queries (This is a 7h span). This is because the media information f...
the network stopped working again
Even with thousands of requests, your network shouldn't just "stop". What router are you using?
well not exactly stop, but being extremely unresponsive on the dns requests
like discord doesnt load images, not even profile pictures as example
Ok, that might be due to throttling from your router or Cloudflare
Can you post onto the above linked issue then? We do not track bugs in Discord.
on your link? can you reopen the issue?
or i just make new?
On the link, yes. I will reopen as soon as you post. The issue is not locked and you should be able to comment.
commented, tried to wrap everything, not sure if i missed something
Hey, any update on this?
Cus right now I really can't have the container running, it's been offline since that day
well something whatever happened
today i launched the container again to see if the issue was randomly fixed
and all my configuration is completely wiped
Please post your docker compose / run config
Or do you run from source?
docker portainer, i just booted the container again
Please post the compose.
is the secret encrypt key safe to show?
No.
You likely upgraded from 0.15.10 incorrectly. Please see https://homarr.dev/docs/getting-started/installation/docker#installation for instructions.
Docker | Homarr documentation
Docker is our recommended installation method for beginners and professionals.
uh?
Migrate from 0.15.10+ to 1.0.0 | Homarr documentation
In this guide, we will show you how to migrate from Homarr 0.15.10+ to 1.0.0.
when was that version released?
What version? 0.15.10?
We released 1.0 in January this year.
Latest is 1.21.0 at the moment.
i did the upgrade
But incorrectly. Hence your data reset.
See the link above for the correct compose 🙂
how is it possible, it was working 4 days ago on 1.20
the container its been offline since then
Because the data is not persisted if you forget to mount directories. It worked, but the files were then lost when the container was destroyed.
im missing a point
no container was destroyed/created
the container is the same, as it was 4 days ago
here you did.
i did the upgrade on 1.0 release day, 4 months ago
not 4 days ago
thats whats confusing me
i also "beta" tested one of the early 1.0 releases
- You ran 0.15.10
- You upgraded to 1.0 and did some changes
- You started the container again (
today i launched the container again to see if the issue was randomly fixed) and since you did not persist the directories, your changes were lost
I assume it shows the onboarding process again, correct?yeah but it doesnt make sense because its not the first time i kill and start the container lol
i did it that day i ran the dns issue and it was just fine (besides the dns issue)
now all of sudden, its not pulling the config
which i havent touched AFTER i did the upgrade, 5 months ago, and the config was also updated
Let's backtrack please. What are you seeing? Logs? Compose file? Timeline of what steps you took? How did you "start" or "stop"?
And to be clear, it is expected that your container will eventually loose the data, if it's unmounted. That is not a Homarr issue and it is clearly documented in the migration guide and the installation documentation.
So if we start from 1.0 day one, back in janaury, its when i upgraded from 0.15.10 to 1.0
followed both the guides you linked, everything worked perfectly
i also remember some integrations werent working, you were aware of that, eventually got fixed with updates, all good
5 months after (4 days ago) i ran into that dns issue, troubleshooting and so on, Homarr was spamming dns requests
so to fix the flood i stopped the container and came here opening this thread as you remember, then github and whatever
SINCE that day the container has been stopped, never ran again
NOW i am launching it again, and here i am, homarr is asking for bootstrap
see where i am getting? i just stopped the container 4 days ago and launched it again now, config was untouched, actually everything is untouched left the same as it was 4 days ago
i stopped the container How did you "stop" it? If Portainer executes a docker compose down, that will delete the unmounted files, which would explain the behaviour you're describing.i am not sure how portainer interacts with it, but it has 2 commands
stop and kill
Yes, I see what you mean. But in our guides we had for 1.0, we showed a different compose than what you have here. Your compose does not mount correctly and this caused your data loss. Sorry that you lost this data but it is not something Homarr can auto-detect.
thats the thing
i didnt lose any config
i am in the config folder with WinSCP and the file is there, last edit 21/05/2025
default.json has everything in iit
then the problem is portainer hereWell, obviously you did. Otherwise, Homarr wouldn't start over.
The directory
/app/data/configs and /app/public/icons are no longer in use in 1.0.
And to be clear, default.json is from your old Homarr (0.15.10), not 1.0. So no, your data is no longer there.
If your default.json has that date, it means that you ran 0.15.10 until 21.05.no my bad, i read the date wrong, its back in 2024
so what is the path of 1,0 config
./homarr/appdata?
Sorry, I don't mean to be offensive, but I've already told you multiple times. Check the docs.
See https://homarr.dev/docs/getting-started/installation/docker#installation
/appdata is what needs to be mounted. It doesn't matter to what directory or volume you mount to on your host!Docker | Homarr documentation
Docker is our recommended installation method for beginners and professionals.
i looked into it, that line is missing from my docker compose but then i cant understand how i was able to run (and update it) for 5 months without any issues
As mentioned as above, It runs without the mount. It will just not persist the data. This is normal and expected.
anyway if
/appdata mount point is missing from the compose, does the image creates it by default somewhere?Docker might have auto-created an anonymous volume somewhere. I can't help you with finding that though. You could check the currently available volumes in the Docker engine.
If you can't find it, no - then your data would be gone
is it a requirement that the appdata path is flagged as hidden?
Flagged as hidden? I don't understand your question.
./homarr/appdata/ doesnt the . stands for hidden?No it's a relative path...
I suggest you to have a look at https://www.youtube.com/watch?v=DM65_JyGxCo&pp=ygUXZG9ja2VyIGNvbXBvc2UgYmVnaW5uZXI%3D
NetworkChuck
YouTube
Docker Compose will BLOW your MIND!! (a tutorial)
We don’t live in Docker containers, keep yourself safe with Bitdefender Premium Security: https://bit.ly/BitdefenderNC (59% discount on a 1 year sub)
Stop spinning up Docker Containers like your grandparents, use Docker Compose instead! No more worrying about typing a million commands to get your whole Docker network and Containers running,...
anyway i cant import the old default.json i have right? i have to start from 0 again
the bad config i have might also be the reason i had the dns issue in first place
You can spin up an oldmarr (0.15.10) instance and add it in your configs directory there
true that would be smart
Then you can proceed with the export in the manage / tools / migrate to 1.0 page and export your board & stuff
where can i find the old image link?
thanks a lot
There is no documentation anymore, but I've went to the docs repository and viewed the old compose there, if something does not work, let me know
should work, i remember very similarity with the config i had
just another question
the
default.json does not contains informations about the admin accounts right?
i am being asked to create new
yep exactly, i have everything nowYes, default.json only contains the board, but you may have an existing db.sqlite file also from the old instance on your data mount.
But you can of course also just recreate the user 😉
yeah i did, the passwords requirements were also different
going to migrate again
this old version image was a nice save into a corner, gg
got homarr back to work but the dns issue persist
infinite spam of dns requests 😦
Glad you got it back. Is the TTL still above 0 and Homarr still shows the same issue?
yes
64
and yes, flood of dns requests over and over
Solution
Ok, then let's just keep tracking this in the GitHub issue as we discussed.
Can I mark this as resolved?
discord thread yeah
ill wait for github updates
Yes, of course. Thank you!
+1
Please post on GitHub too with as much detail as possible