UniFi: ctrld tries mongodb, even though mongodb does not exist

[dcabri]

#OS Type : linux
#OS Vendor : ubios
#OS Version : 5.0.12.30269
#Router Model : Gateway Max
#Arch : aarch64
#CPU : Kryo V2
#Free RAM : 750 MB / 1946 MB
#Binary URL : https://dl.controld.com/linux-arm64/ctrld
#Install Path : /data/controld
ctrld version v1.4.9

ctrld is trying to launch mongodb, even though this is not present:

/bin/sh /usr/bin/mongo localhost:27117/ace --quiet --eval
DBQuery.shellBatchSize = 256;
db.user.find({name: {$exists: true, $ne: ""}}, {_id:0, mac:1, name:1});

root@Max:/data/controld# systemctl status ctrld
● ctrld.service - A highly configurable, multi-protocol DNS forwarding proxy
Loaded: loaded (/etc/systemd/system/ctrld.service; enabled; vendor preset: enabled)
Active: active (running) since Wed 2026-02-18 08:33:53 EET; 2s ago
Main PID: 1987116 (ctrld)
Tasks: 14 (limit: 2327)
Memory: 36.1M
CPU: 1.503s
CGroup: /system.slice/ctrld.service
├─1767778 /bin/sh /usr/bin/mongo localhost:27117/ace --quiet --eval
DBQuery.shellBatchSize = 256;
db.user.find({name: {$exists: true, $ne: ""}}, {_id:0, mac:1, name:1});
├─1767779 sleep 99999
└─1987116 /data/controld/ctrld run --cd xxxxxxxxxxxxx --iface=auto --homedir=/etc/controld --config=/etc/controld/ctrld.toml

Feb 18 08:33:53 Max systemd[1]: Started A highly configurable, multi-protocol DNS forwarding proxy.
Feb 18 08:33:54 Max ctrld[1987116]: Feb 18 08:33:54.518 NTC internal logging enabled
Feb 18 08:33:55 Max ctrld[1987116]: Feb 18 08:33:55.467 WRN failed to init Ubios discover error="out: , err: fork/exec /usr/bin/mongo: no such file or directory"

[matrix8967]

Thanks for reaching out. It looks like the ctrld service is up and running despite not being able to initialize the mongodb requirement.

Is ctrld running and handling DNS as expected, despite the mongodb error? (You can check by running sudo ctrld status -vv and nslookup verify.controld.com, as well as checking your Control D dashboard.)

[dcabri]

I personally think - as do the AI engines - this error occurs for every single lookup for a newly discovered client.
Is it working? yes
Is it significantly slowing down processing? yes.

I have a UDR and a Gateway Max.

About the UDR:

• has no mongodb
• is a cloud router
• no mongodb error shows

About the Gateway Max

• has no mongodb
• is NOT a cloud router
• mongodb error shows

UDR: no issues with responding to DNS requests.
Gateway Max: very slow in responding to DNS requests.

To satisfy the ctrld daemon, I created a shell script called /usr/bin/mongodb, containing only 'sleep 9999'
This had as a result that NO DNS processing was done. The crtld daemon was completely blocked.
Hence, the DNS processing is checking mongodb for every DNS queuery.
If it does not exist, this induces latency.

I do not expect the simple rule:
"if mongodb executable does not exist or is not executable, skip any mongodb processing"
to be very hard to implement. Hence the request.

My conclusion is that the ctrld software is wrongly identifying and understanding the Gateway Max UniFi devices, and that it seems unable to deal with the absence of mongodb executable at the moment.

As to answer your questions:

root@Max:/data/controld# sudo ./ctrld status -vv
Feb 18 20:30:44.000 NTC Service is running

root@Max:/data/controld# nslookup verify.controld.com
Server: 127.0.0.1
Address: 127.0.0.1#53

Non-authoritative answer:
verify.controld.com canonical name = api.controld.com.
Name: api.controld.com
Address: 147.185.34.1
Name: api.controld.com
Address: 2606:1a40:3::1

[matrix8967]

Gotcha -- running, albeit not very well. :/

Thanks for the report & the detailed follow up. I've got this added to our board for the dev team.

In the mean time - you can probably uninstall ctrld with sudo ctrld uninstall --cleanup -vv to remove the service and run ctrld in ephemeral mode while we work on a permanent fix.

Below is an example script to download and run ctrld in ephemeral mode:

#!/bin/sh

export PATH=$PATH:/data/controld

wget --content-disposition https://dl.controld.com/linux-arm64/ctrld

chmod +x ctrld

./ctrld run --listen=127.0.0.1:5354 --primary_upstream=https://dns.controld.com/$RESOLVER-ID-HERE --log=/data/controld/ctrld.log -vv

If this works - the script could be improved by prefixing the ctrld run... with a nohup and sending the process to the background with & which will leave the process running in the background without requiring login:

nohup ./ctrld run --listen=127.0.0.1:5354 --primary_upstream=https://dns.controld.com/$RESOLVER-ID-HERE --log=/data/controld/ctrld.log -vv &

^^ This isn't a permanent or long-term solution, but it may help alleviate the performance issues on your device.

edit As you can see below -- even ephemeral mode still causes this issue. :/

[dcabri]

Thank you so kindly for the help.

I did waht you asked and now get confirmation:

{"level":"warn","error":"out: , err: fork/exec /usr/bin/mongo: no such file or directory","time":"2026-02-18T22:44:53+02:00.447","message":"failed to init Ubios discover"}
{"level":"debug","error":"exec: "avahi-browse": executable file not found in $PATH","time":"2026-02-18T22:44:53+02:00.483","message":"could not find avahi-browse binary, skipping."}

Something's wrong:
root@Max:/var/log# grep 'found hostname' controld.log | wc -l
200
root@Max:/var/log# grep 'update hostname' controld.log | wc -l
127794
... and that number keeps climbing.

ctrld log -rw------- 1 root root 27M Feb 18 22:52 controld.log
after only 8 minutes.

Switching -vv off again...

Thanks again!

[matrix8967]

Thank you so kindly for the help.

I did waht you asked and now get confirmation:

{"level":"warn","error":"out: , err: fork/exec /usr/bin/mongo: no such file or directory","time":"2026-02-18T22:44:53+02:00.447","message":"failed to init Ubios discover"} {"level":"debug","error":"exec: "avahi-browse": executable file not found in $PATH","time":"2026-02-18T22:44:53+02:00.483","message":"could not find avahi-browse binary, skipping."}

Something's wrong: root@Max:/var/log# grep 'found hostname' controld.log | wc -l 200 root@Max:/var/log# grep 'update hostname' controld.log | wc -l 127794 ... and that number keeps climbing.

ctrld log -rw------- 1 root root 27M Feb 18 22:52 controld.log after only 8 minutes.

Switching -vv off again...

Thanks again!

Gotcha -- so even running in ephemeral mode w/o the service/installation causes this. 😔 What a bummer.

If it's running in the background - you can kill the background service w/ sudo kill $(pidof ctrld).

You could disable client discovery completely - I'm not sure if only disabling ubios client discovery is a valid config option. 🤔

I'll defer to @cuonglm 's wisdom here.

[dcabri]

JOY!!

It stopped sending megabytes to you, AND did not look for monogdb anymore!
I used
discover_hosts = false

Thank you SO SO much!!!

[matrix8967]

JOY!!

It stopped sending megabytes to you, AND did not look for monogdb anymore! I used discover_hosts = false

Thank you SO SO much!!!

Literally happy to help. Thanks for reaching out and doing the detective work. 🕵️

I'll leave this issue open for any other users who run into this.

Let us know if you run into anything else - otherwise we'll have a fix out soon™️.

[cuonglm]

I personally think - as do the AI engines - this error occurs for every single lookup for a newly discovered client.
Is it working? yes
Is it significantly slowing down processing? yes.

Nope,this only run once during initialization, so no slowing down when new clients discovered.

[dcabri]

Hello,
Thank you for your comment. Unfortunately, in the above we were able to prove your statement is (at least partially) incorrect, as it generates an impressive log file of 27Mb in only 8 minutes, repeating “found” and “sending” “new client” over and over again (any client is always a “new” client) to the Control-D servers. So not having mongodb access creates a huge overhead.
Sorry.
But I am very grateful for the comments from @matrix8967 that has given a way to get around the issue.
Thank you for your comment anyways