Statistics are reset after each reboot

[ammnt]

Prerequisites

Please answer the following questions for yourself before submitting an issue. YOU MAY DELETE THE PREREQUISITES SECTION.

• I am running the latest version
• I checked the documentation and found no answer
• I checked to make sure that this issue has not already been filed

Issue Details

• Version of AdGuard Home server:
  • 0.105.0
• How did you install AdGuard Home:
  • Github releases
• If it's a router or IoT, please write device model:
  • VPS
• Operating system and version:
  • Debian 10

Expected Behavior

Statistics are saved after reboot.

Actual Behavior

Statistics are reset after each reboot actually.

Additional Information

The logs is attached.

[ainar-g]

Hi, can you please show, what your statistics interval is set to? It's the statistics_interval setting in the configuration file.

[ammnt]

@ainar-g, statistics_interval: 30. The difference in this installation was that I install the server under an unprivileged user via sudo. Possible need to check the installation script... It seemed to me that I had already created such a problem before, but I can not find it now in repo😪

[ainar-g]

The current stats code seems to hide a lot of useful information, unfortunately. I've just pushed a623ac6, where I added more verbose logging to the code. Could you please update and collect the new logs? Thanks!

[ammnt]

@ainar-g, unfortunately, that server is gone, but i tried to reproduce it on my VM and it looks fine now🤷🏻‍♂️ IDK how to help here🙄

[ammnt]

@ainar-g, i found it:
#1785
May be useful😒

[ammnt]

@ainar-g, I have successfully reproduced the issue again on the latest edge version. The log attached - please, take a look. The also screenshots is as confirmation is attached:


The second screenshot - its after reboot. Thank you.

[ainar-g]

That's not statistics though, that's query log.

How exactly do you reboot the machine? AdGuard Home keeps a buffer of recent queries in the RAM, so if it is not shut down properly, that buffer could remain unflushed to the filesystem.

[ammnt]

@ainar-g, via my hosting provider panel and this was not before on 0.102.0, 0.103.0 etc. version🤨

[ainar-g]

I don't think that that particular mechanism has changed much in the recent releases. At least I don't remember that. If you can, you should set up your machine to kill the AGH process with SIGINT, which should force flushing, and wait until it goes down. Alternatively, you could also set dns.querylog_size_memory to a lower value, so that the flushing happens more often.

[ammnt]

@ainar-g, i set querylog_size_memory 10 instead 1000 and restart the server. All query logs and statistics are gone after that😂

[ainar-g]

Are you sure that your data/querylog.json is not deleted by some other entity? That's the file where the logs are kept. And the statistics data is also in the data/ directory.

[ammnt]

@ainar-g, yes, i'm sure. Moreover, part of the statistics disappears, and not all.

[ainar-g]

I have no ideas besides the VPS's filesystem shenanigans. We've never seen anything similar. If you have a stable reproduction on a machine that is not on the same VPS, please let us know.