Open Ice2192 opened 1 year ago
Same here, it starts well and suddenly all devices cannot reach internet. I had to restore previous version backup (v0.107.28) and all was fine since then.
Same here. Running in Docker container on Synology NAS.
I've pinned the Docker image version to v0.107.29
to get it working again.
I've noticed in the Query logs that the DNS queries of my computer were shown but the Elapsed time was very high. They were in the range of 4300 ms or 7500 ms. This was probably why the DNS requests were timing out and not working anymore. It seems only to be having issues with forwarding DNS requests to the upstream DNS servers as DNS rewrites were working fine.
Same problem. My watchtower did an update last night, basically no internet connection since then. Reverted back to 0.107.29, problems gone.
same problem @ameshkov @ainar-g need help
Hello @Ice2192, @werdlerk, @dfsworld, @andreschoppe, @Insight16! Thanks for reporting. Could you please collect a verbose log capturing the incoming requests? This would help us troubleshoot this issue. You may send it to devteam@adguard.com.
This is likely caused by the same bug as #5873, so the fix for that should also improve the upstream situation generally.
@Ice2192, @werdlerk, @dfsworld, @andreschoppe, @Insight16, we've just pushed the edge
build that should fix this. Could you please check if it now works as intended? If it does, we're going to prepare a proper hot-fix release.
Same problem. My watchtower did an update last night, basically no internet connection since then. Reverted back to 0.107.29, problems gone.
How do you downgrade the version via SSH?
@Ice2192, @werdlerk, @dfsworld, @andreschoppe, @Insight16, we've just pushed the
edge
build that should fix this. Could you please check if it now works as intended? If it does, we're going to prepare a proper hot-fix release.
Can confirm edge build works as intended. No more problems here. Thanks guys.
Same problem. My watchtower did an update last night, basically no internet connection since then. Reverted back to 0.107.29, problems gone.
How do you downgrade the version via SSH?
You can always use any version with docker that is still available in the hub by using the tags. E.g. docker run --name adguardhome --restart unless-stopped -v /pathtoyouroptdir:/opt/adguardhome/work -v /pathtoyourconfdir:/opt/adguardhome/conf -p 8080:80/tcp -p 3000:3000/tcp -p 53:53/tcp -p 53:53/udp -d adguard/adguardhome:v0.107.29 You can find all available tags on the docker hub page for adguard
@Ice2192, the newly released v0.107.31 should contain a fix. Please update to it and tell us if the slowdown went away.
@Ice2192, the newly released v0.107.31 should contain a fix. Please update to it and tell us if the slowdown went away.
Thanks! It's back to its normal speed!
I'm going to test it later but many thanks for the quick response and resolution! 👍
No dice. It doesn't work anymore. Worked for about an hour and it's back to slowdown again.
No dice. It doesn't work anymore. Worked for about an hour and it's back to slowdown again.
Same for me, it worked at first, but I still get a boat load of errors in the interface and DNS timeouts
Same for me. Worked fine for the first couple of hour, the started slowing down and then failing entirely.
Edge version still doing fine for me ...
Edge version still has issues.
@Ice2192, @Xerdunits, @degofedal, @andreschoppe, @Mur1thi, thanks for your concern. Could you please collect a verbose log again, so that we could revise the actual source of the current problem? You may, again, send it to devteam@adguard.com.
I confirm that I had to pin v0.107.29
to avoid random connection drops due to query failures. I'm running it on ARM64 and ARMv7, via Docker, and on both architectures was causing issues with v0.107.30
and v0.107.31
.
You can also see how my HomeAssistant instance wasn't able to call the API sometimes from this graph showing missing data from both instances (AdguardHome runs in the same host so no network issue here but API issue).
Same problem. My watchtower did an update last night, basically no internet connection since then. Reverted back to 0.107.29, problems gone.
How do you downgrade the version via SSH?
You can always use any version with docker that is still available in the hub by using the tags. E.g. docker run --name adguardhome --restart unless-stopped -v /pathtoyouroptdir:/opt/adguardhome/work -v /pathtoyourconfdir:/opt/adguardhome/conf -p 8080:80/tcp -p 3000:3000/tcp -p 53:53/tcp -p 53:53/udp -d adguard/adguardhome:v0.107.29 You can find all available tags on the docker hub page for adguard
Sorry I didn't use docker to install AGH on my Pi. I just typed in a line of code that someone said on YouTube which immediately took the latest version available. Is there a similar one one line of code that will work to just install XX.29 on the pi without resetting the settings and query history?
@Xerdunits, hello. I've noticed the usage of DNSCrypt upstream through your log. Could you please try the latest beta? It contains some extra DNSCrypt-related fixes, that may affect your AGH instance. We're currently actively working on the issue, your feedback helps a lot.
@Ice2192, @degofedal, @andreschoppe, @Mur1thi, please, consider trying beta too if you're using DNSCrypt.
@Xerdunits, yeah, I'd suggest you to send it via email to devteam@adguard.com.
@Xerdunits, well, it's indeed broken. We're keep working on it and going to push new fixes later today, wouldn't you mind to check those? I suppose, your particular case is still related to DNSCrypt, but for the moment let's keep it here.
@Xerdunits, yes, but again, consider sending it to devteam@adguard.com, since verbose logs are usually quite massive, navigating the issue becomes difficult.
So far it works! I'll let this thread sit at open for 24 hours if that's okay in case things change in the meantime.
I’m still having challenges, even after updating.
So far it works! I'll let this thread sit at open for 24 hours if that's okay in case things change in the meantime.
Yep. Been using the AGH since then and about 2 hours ago when I'm typing this is when it slowed down again.
Just an update from my side - I am using the edge build since last week. No more issues since then ...
So far it works! I'll let this thread sit at open for 24 hours if that's okay in case things change in the meantime.
Yep. Been using the AGH since then and about 2 hours ago when I'm typing this is when it slowed down again.
Since I'm using a Pi as well, what is your setup? Do you use multiple DNS server? When you say 'slowed down', that means it is semi working and DNS resolves take longer than usual or does it stop entirely?
Pi as well. I just use the default DNS servers that appear when you set up AGH. By slowed down I mean that when I go to youtube only half the thumbnails load and then stays at loading. Speedtest app for Windows doesnt even want to start.
@Ice2192, it seems, your case isn't related to DNSCrypt. Could you please collect the verbose log and send it to devteam@adguard.com? This would really helps us to troubleshoot this issue.
So far it works! I'll let this thread sit at open for 24 hours if that's okay in case things change in the meantime.
Yep. Been using the AGH since then and about 2 hours ago when I'm typing this is when it slowed down again.
Since I'm using a Pi as well, what is your setup? Do you use multiple DNS server? When you say 'slowed down', that means it is semi working and DNS resolves take longer than usual or does it stop entirely?
Pi as well. I just use the default DNS servers that appear when you set up AGH. By slowed down I mean that when I go to youtube only half the thumbnails load and then stays at loading. Speedtest app for Windows doesnt even want to start.
So you use the same DNS IP address when you manually set it on your computer, or a different server?
I ask my router to look at my Pi for the adblock. Nothing changes on any of my devices. When a slowdown like the 30 update came a long and I diagnosed that the slowdown was coming from the Pi, I changed the DNS servers on my router to Cloudflare temporarily until there is a fix.
@Ice2192, it seems, your case isn't related to DNSCrypt. Could you please collect the verbose log and send it to devteam@adguard.com? This would really helps us to troubleshoot this issue.
Sorry, still sort of a noob at this. I opened Filezilla to get the yaml file and made the edits to it as requested. I don't know how to stop adguard home. I typed
AdGuardHome -s stop
but the command line says "command not found"
@Ice2192, the command AdGuardHome -s stop
is only applicable if AdGuard Home have been installed as a service. Have you used an installation script from README.md
or have you downloaded the binary from releases page?
@EugeneOne1 I have used a script to install.
@Ice2192, in order to call the service commands, the full path to AdGuard Home executable should be specified. As far as I remember, on Windows it should be somewhere about the root of C:
drive, but I could be wrong since the script doesn't officially support Windows. You may actually figure out the path to the executable with Windows task manager through the AGH's process.
I'd also recommend you to use an absolute path for the log_file
configuration property.
@EugeneOne1 The executable is not on windows but on the pi itself. I found the folder of AG is in the pi but I dont see the executable. Do you know the default location of it. When I installed AGH I just copied and pasted the script onto windows shell to ssh into the pi.
@Ice2192, oh, I got it. On Unix systems the script puts the binary under /opt/AdGuardHome
, the binary itself also called AdGuardHome
, so you should be able to call it with something like:
/opt/AdGuardHome/AdGuardHome -s stop
@EugeneOne1 Yeah. I tried that and I got this:
[fatal] service: executing action "stop": Failed to stop AdGuard Home service: exit status 1
@Ice2192, if the AGH itself still running? The command
ps ax | grep AdGuardHome
will print two processes if it is. If so, I'm sure it would be safe to kill the AGH's process with kill NNNNN
.
@EugeneOne1
315 ? Ssl 5:02 /opt/AdGuardHome/AdGuardHome -s run
764 pts/0 S+ 0:00 grep --color=auto AdGuardHome
If you installed it via the install script, you probably also have a systemd service. Try:
sudo systemctl stop AdGuardHome.service
Thanks! It did stop AGH however when I tried to copy the yaml file with the verbose edit via filezilla, I gave me a permissions error. I don't want to waste even more of your time, as you've been very patient with me thus far, and ask how to remove it so I pasted it the old fashioned way of turning off the pi and taking out the sd card and paste it via my nas as it's the only other way i could view the files.
Do I just connect devices to the AGH and visit websites for the thing to input information into the log and then take out the log and email you guys?
@Ice2192, that's right. It would be great, if the actual issue case appeared and have been captured. Thanks for your engagement.
@EugeneOne1 I sent it.
@Ice2192, hello again. We've inspected your log and unfortunately found nothing suspicious. Could you please provide some details:
It also may be related to the filtering. Could you please try to disable the protection for some time and check if the problem persists? And also, could you please filter the query log by blocked status and check if something legal have been blocked there?
@EugeneOne1 Any app / website doesn't work. If I go to any website, it just either hangs in the loading screen or I get an error as mentioned in the first post. It was working fine until .30 update came along. Haven't changed any blocklist since then. I usually get a ping of 2 but the update causes the ping to be ≥ 5. I used to use a microcenter sd card that they gave away for free but those causes the pi to hang when it needed to update the blocklist which I would have to force restart the pi everytime. Then I got a san disk one and since then haven't had problems until now. Hard resetting to clear cache for both my routers still didn't change a thing. Right now every time I point my router to the pi it works for like an hour or 2 and then just slows to a halt.
@EugeneOne1 Is there a way to downgrade to 29. Anything after that version completely made everything slow. I even did a clean wipe of the sd card and reinstalled Pi OS and AGH .
FYI: having the same issue - all was well with v29 and ever since then I've had to turn off AGH because it noticeably slows down traffic. on v32 all seems well for the first few minutes and then eventually crawls. I uninstalled/re-installed through homebridge with same result.
FYI: having the same issue - all was well with v29 and ever since then I've had to turn off AGH because it noticeably slows down traffic. on v32 all seems well for the first few minutes and then eventually crawls. I uninstalled/re-installed through homebridge with same result.
@boarder010 Did you disable only the AGH protection from the button in the main page or the entire service switching to different DNS?
Experiencing the same with the latest version too. AdGuard home works well in the first few minutes then it breaks and I'm no longer connected to the internet.
Prerequisites
[X] I have checked the Wiki and Discussions and found no answer
[X] I have searched other issues and found no duplicates
[X] I want to report a bug and not ask a question
Operating system type
Windows
CPU architecture
64-bit ARM
Installation
GitHub releases or script from README
Setup
On a router, DHCP is handled by the router
AdGuard Home version
v0.107.30
Description
What did you do?
Go on the internet.
Expected result
View pages.
Actual result
Error: "This Site Can’t Be Reached"
Screenshots (if applicable)
Additional information
I just updated ASAP when I got the notification today and then the internet just drops. I thought it was the router but I changed the DNS on my PC and it was fine. I recently upgraded from a miro center sd card to a sandisk a few months ago as the microcenter one hangs when updating the blocklists for some reason. Had no problems since the san disk upgrade. Now with this update, the DNS just gets lost when I point anything to my pi running adguard home.