Closed z1ucas closed 6 months ago
@z1ucas Thanks for reporting.
Does /data/unifi
exist in your system?
@cuonglm No that doesn't exist. Below is the contents of the /data directory :
controld/ dpi-tracer/ udapi-config/ ui-db/ uxgpro-setup/ wifiman/
@cuonglm No that doesn't exist. Below is the contents of the /data directory :
controld/ dpi-tracer/ udapi-config/ ui-db/ uxgpro-setup/ wifiman/
Thanks, that explains the failure.
ctrld
is checking for /data/unifi
to detect it's running on Ubios.
Probably we need a better way for this checking.
ctrld
is checking for/data/unifi
to detect it's running on Ubios.Probably we need a better way for this checking.
Ok, So I just created the /data/unifi
directory and its now installed successfully.
@z1ucas is this a fresh installation of UniFi UXG?
@z1ucas What does this command return? ubnt-device-info summary
@yegors
Device information summary: Subsystem ID: ea19 Family: UniFi NeXt-Gen Gateway (UXG) Model: UniFi NeXt-Gen Gateway Pro (UXG-Pro) Default MAC address: XX:XX:XX:XX:XX:XX Default IPv4 address: 127.0.0.1 Firmware: 3.1.16 (3.1.16)
Thanks!
@cuonglm ^ thats your anchor
Thanks!
@cuonglm ^ thats your anchor
We can't use ubnt-device-info summary
in ctrld
, because that binary is not available during boot time when ctrld was started by system service manager.
The installer can use it, because the installer is run after boot.
@yegors @cuonglm FYI I have a UXG Lite as well and the same issue existed. To get past it in the interim I added the /data/unifi
directory :
Device information summary: Subsystem ID: a677 Family: UniFi NeXt-Gen Gateway (UXG) Model: UniFi NeXt-Gen Gateway Lite (UXG) Default MAC address: XX:XX:XX:XX:XX:XX Default IPv4 address: 127.0.0.1 Firmware: 3.2.11 (3.2.11)
@z1ucas Could you please try .dev domain and see if the issue was fixed?
Thank you.
This should be fixed after #138
Apologies, I’ve only just got to this message. I’ll also check it out and install on a UXG device. If there are any problems I’ll reply back.On 5 Mar 2024, at 04:37, Cuong Manh Le @.***> wrote: This should be fixed after #138
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>
@z1ucas I assume all good?
No, I’m having a few problems on both UXG Lite and I’m having intermittent DNS issues on the UXG Pro. DNS seems to stop responding on both UXG Lite requiring the daemon to be shut down. Upon restarting the daemon it’s instantly unresponsive.I’m trying to pinpoint what is going on. I’ll post back errors that I’m seeing here in about 15 mins.On 9 Mar 2024, at 03:46, Yegor S @.***> wrote: @z1ucas I assume all good?
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>
That doesn't seem related to the issue thread. If you have other problems, open a new issue and collect debugging data while it's in the broken state: https://github.com/Control-D-Inc/ctrld/wiki/Troubleshooting-Guide
That doesn't seem related to the issue thread. If you have other problems, open a new issue and collect debugging data while it's in the broken state: https://github.com/Control-D-Inc/ctrld/wiki/Troubleshooting-Guide
I am pretty sure it is related. I have been trying to work out why DOH, DOH3 and DOQ stopped working on all my PRO and LITE UXG devices when upgrading. The first error that I am seeing is :
Mar 9 09:03:31.988 ERR could not init Ubios discover error="fork/exec /usr/bin/mongo: no such file or directory"
It does look related to my initial query in relation to detecting what the device is.
After that, a range of error messages appear, depending on what you try to do with the config. I should note that it is impossible to install using a config hosted at ControlD, the following error occurs :
Mar 9 01:21:01.968 ??? ================================
Mar 9 01:21:01.969 ??? An error occurred while performing test query: read udp 127.0.0.1:46581->127.0.0.1:5354: i/o timeout
Mar 9 01:21:01.969 ??? ================================
Mar 9 01:21:01.970 ??? ctrld service was running, but a DNS query could not be sent to its listener
Mar 9 01:21:01.970 ??? Please check your system firewall if it is configured to block/intercept/redirect DNS queries
Mar 9 01:21:01.970 ??? ================================
Mar 9 01:21:17.214 NTC Service uninstalled
Once I take the exact same config hosted at ControlD it's possible to start it locally with DOQ, but DOH and DOH3 must be removed from the config or it won't start. Even though it starts with DOQ - no DNS queries are successful.
Also, during the install the following is being displayed, which is different than what we've seen in the previous builds. Not sure if this makes a difference. Lite then Pro :
---------------------
| System Info |
---------------------
OS Type : linux
OS Vendor : ubios
OS Version : 3.2.15.14839
Router Model : Gateway Lite
Arch : aarch64
CPU : Kryo V2
Free RAM : 178 MB / 974 MB
---------------------
| Install Details |
---------------------
Resolver ID : #REMOVED#
Binary URL : https://dl.controld.com/linux-arm64/ctrld
Install Path : /data/controld
---------------------
| System Info |
---------------------
OS Type : linux
OS Vendor : ubios
OS Version : 3.2.15.14839
Router Model : Gateway Pro
Arch : aarch64
CPU : Cortex-A57
Free RAM : 141 MB / 1997 MB
---------------------
| Install Details |
---------------------
Resolver ID : #REMOVED#
Binary URL : https://dl.controld.com/linux-arm64/ctrld
Install Path : /data/controld
If you created /data/unifi, does v1.3.4 binary work for you?
Could you please run ctrld interactively using ctrld run --cd=<uid> -vv
, then sending query directly to ctrld listener.
So, I did find the culprit and I’m not sure why it has a range of issues associated with it.
I have UniFi Magic Site active across 3 locations. After the upgrade to either 1.3.4 or 1.3.5 the following line which is referenced from the configuration docs causes Site 2 and Site 3 to stop working (if it was running) or fail to start CtrlD if it wasn’t running / or being installed - when Site 1 has CtrlD running or is commanded to start up. Site 2 and 3 are behind a firewall/router and Site 1 has the Public IP :
An empty upstream would not route the request to any defined upstreams, and use the OS default resolver.
[listener.0.policy]
name = "OS Resolver"
rules = [
{"*.local" = []},
]
I am not 100% sure what it is doing, but the .local rule is definitely causing a range of different issues.
Site 2 and 3 work perfectly if Site 1 doesn’t have CtrlD running, but as soon as Site 1 tries to start CtrldD with this rule all 3 sites begin failing.
Removing the rule from Site 1 allowed all sites to start and run CtrlD, though I removed it from all 3 sites in any case.
Hi Guys,
I am seeing the following error when trying to install on a Unifi UXG Pro, is there any clues as to how this can be resolved?
The device does have 2 WAN interfaces and an additional Backup 5G WAN.