[BUG] The http.conf file is not written after modifying the service configuration in the Web UI.

[weizhuangzhi]

[BUG] The http.conf file is not written after modifying the service configuration in the Web UI.

[TheophileDiot]

Hi @weizhuangzhi, I couldn't manage to reproduce your error. Is it possible that you give us more information about the steps to reproduce your error ?

[weizhuangzhi]

Could it be related to url accessibility in Global_config's DNSBL list?

Sometimes these URLs cannot be accessed in an intranet environment, and Plugins sometimes displays: PLUGINS ERRORS 1

Can it be reproduced if the URL in the DNSBL list is set to a non-existent URL?
If this affects whether the configuration can be saved, ignore this accessibility issue. I have tested it many times and it seems that it cannot be reproduced every time. If a problem occurs and it seems that the program has interrupted execution, delete the configuration directory of the site and execute this process. But the re-creation process is not executed.

[weizhuangzhi]

Reinstalled and tested again.

sudo dnf update -y
sudo dnf install -y nginx-1.24.0
curl -s https://packagecloud.io/install/repositories/bunkerity/bunkerweb/script.rpm.sh | \
sed 's/yum install -y pygpgme --disablerepo='\''bunkerity_bunkerweb'\''/yum install -y python-gnupg/g' | \
sed 's/pypgpme_check=`rpm -qa | grep -qw pygpgme`/python-gnupg_check=`rpm -qa | grep -qw python-gnupg`/g' | sudo bash && \
sudo dnf makecache && \
sudo dnf install -y bunkerweb-1.5.2

sudo tee /etc/bunkerweb/ui.env >/dev/null <<EOF
ADMIN_USERNAME=administrator
ADMIN_PASSWORD=*************
EOF

sudo tee /etc/bunkerweb/variables.env >/dev/null <<EOF
DNS_RESOLVERS=8.8.8.8 8.8.4.4
HTTP_PORT=80
HTTPS_PORT=443
API_LISTEN_IP=127.0.0.1
API_HTTP_PORT=7001
SERVER_NAME=bunker6.tt
MULTISITE=yes

bunker6.tt_USE_UI=yes
bunker6.tt_USE_REVERSE_PROXY=yes
bunker6.tt_REVERSE_PROXY_URL=/manage
bunker6.tt_REVERSE_PROXY_HOST=http://127.0.0.1:7000
bunker6.tt_INTERCEPTED_ERROR_CODES=400 404 405 413 429 500 501 502 503 504
EOF

# restart service
sudo systemctl restart bunkerweb
sudo systemctl status bunkerweb --no-pager
sudo systemctl restart bunkerweb-ui
sudo systemctl status bunkerweb-ui --no-pager

# open firewall web ports
sudo firewall-cmd --permanent --add-service=http
sudo firewall-cmd --permanent --add-service=https
sudo firewall-cmd --reload

Visit http://bunker6.tt/manage/ to enter the Web UI. In order to simulate the situation where the DNSBL URL is inaccessible, set the Global configs DNSBL list to: dnsbl.unknown.net, and then save it.

To avoid the impact of caching, etc., run the following command:
sudo systemctl restart bunkerweb
sudo systemctl restart bunkerweb-ui

Enter the Web UI, create a service, set the Server name to: mysite.net, save other settings without setting them, and view the files in the /etc/nginx directory:

Enter the Web UI, edit the mysite.net service, save it directly without modifying any settings, and view the files in the /etc/nginx directory:

It is not a problem of slow execution of the background program, because after waiting for a long time, there is still no mysite.net directory generated.

[TheophileDiot]

Hi @weizhuangzhi, we'll shortly push a new testing release for Linux packages, could you test if the issue still occurs with it ?

[weizhuangzhi]

Okay, I'll test it as soon as it's released.

[TheophileDiot]

Hi @weizhuangzhi the dev finally passed (after a lot of tinkering), you can test it on this release if you want

[weizhuangzhi]

I updated the file using the dev branch.

This problem could always be reproduced before, but now it occurs occasionally.

The other test environments have not changed. Edit the service (modify or not modify the configuration items) and save it. About 5-10 times, the mysite.net folder will not be generated, and there seems to be no pattern. (Deleted, recreated, modified many times)

I use watch -n 0.1 -d "ls -l /etc/nginx" on the command line to monitor whether the folder is generated.

In addition, after modifying the service, I encountered "err_connection_refused" errors 2 to 3 times, and the web ui site was also inaccessible. No abnormalities were found using "systemctl status bunkerweb" and "systemctl status bunkerweb-ui". I think the generated nginx configuration may be problems, use "nginx -t":

nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
nginx: configuration file /etc/nginx/nginx.conf test is successful

No configuration error was found, so nginx -s reload was executed.

nginx: [error] invalid PID number "" in "/var/run/bunkerweb/nginx.pid"

It was found that the nginx.pid file was empty and had no pid number. However, this problem is more difficult to reproduce. It may not be a problem with bunkerweb but a problem with my environment. This is for reference only.

[weizhuangzhi]

Comparison of nginx configuration directories when generating files successfully and unsuccessfully:

http.conf : There is one line of difference.

variables.env : All variable definitions starting with "mysite.net_" are missing.

[TheophileDiot]

Interesting, i'll try this again when I have the time

[TheophileDiot]

Hi @weizhuangzhi a few fixes and optimisation have been done since yesterday. If you have the time, you can test the latests changes with the dev tag.
This commit might have fixed your issue : 48096d7

[weizhuangzhi]

@TheophileDiot Thank you, this problem no longer exists in the dev branch.