Cockpit Web Proxy UI

Hi folks,
we started working on the Web Proxy and Filtering UI.

What would you expect from the new UI? What information shall we display in the dashboard and inside other modules?

This is a preview, you can find a working RPM here along with the issue and development wiki page.

Dashboard

Proxy rules

Proxy configuiration

) Screenshot%20from%202019-04-23%2015-39-31

/cc @edoardo_spadoni @davide_marini

9 Likes

just to be sure, the latest packages to install to test it are:
http://packages.nethserver.org/nethserver/7.6.1810/autobuild/x86_64/Packages/nethserver-squid-1.7.4-1.44.pr29.g8745cb0.ns7.noarch.rpm
and
http://packages.nethserver.org/nethserver/7.6.1810/testing/x86_64/Packages/nethserver-squidguard-1.8.0-1.2.gaf019f5.ns7.noarch.rpm

??

Yes, it’s right.

Everything should lend soon on testing, we are just waiting some review from Davide and Filippo.

1 Like

All RPMS have been released!

4 Likes

installed yesterday from testing just before release… playing with it right now, tnx for the great work :clap::clap::vulcan_salute:

3 Likes

Found a little thing:

if antivirus is enabled in cockpit-gui, the c-icap-service is started and enabled, but the clamd@squidclamav is only started not enabled:


I had to enable it manually.

1 Like

i have some problems with web filtering.

  • some filter profiles are marked with the alert icon (this profile use invalid categories), but there are only some categories from shalla (and the global black and whitelist enabled)
  • whenever i change something in a filter profile, ufdb stop to work, restart of the service with systemctl didn’t help, i must disable and then re-enable the filtering service, and also if under Configuration i see blacklist and whitelist enable i must reconfigure the default profile and enable them. After that, all seems to work

i’ll do some other test during the non-working time…

Thank you @flatspin and @dz00te continue the testing session, we will try to fix everything at once! :wink:

1 Like

Fixed in nethserver-squid-1.8.1-1.ns7.noarch.rpm

This is because shalla probably changed the name of some categories. Nothing to worry about, the filter is still working.

I couldn’t reproduce it. :frowning:
@davide_marini could you try to reproduce on your machine?

1 Like

I may have seen the same behaviour in one system that had a lot of filtering profiles.
udfb service restart script left a running ufdbguardd process that I had to kill.
Could you please check that you have a running ufdbguardd process but the /tmp/ufdbguardd-03977 socket is missing?
Then, kill the process and run systemctl restart ufdb

3 Likes

tnx. in effect this is exactly what i see, but also after kill and restart it didn’t work.
but at least this time I had time to read the logs :slight_smile:
it seems that fail to reload whitelist/blacklist/custom, some fatal errors:

2019-06-11 15:09:38 [26830] loading URL table from "/var/squidGuard/blacklists/recreation/restaurants/domains"
2019-06-11 15:09:38 [26830] 
2019-06-11 15:09:38 [26830]    FATAL ERROR: cannot read from "/var/squidGuard/blacklists/custom/whitelist/domains.ufdb" (read-only): No such file or directory  *****
2019-06-11 15:09:38 [26830]     
2019-06-11 15:09:38 [26830] loading URL table from "/var/squidGuard/blacklists/custom/whitelist/domains"
2019-06-11 15:09:38 [26830] ERROR: URL database table "/var/squidGuard/blacklists/custom/whitelist/domains" is empty and is ignored   *****
2019-06-11 15:09:38 [26830] loading URL category defaulttable with creation date 20190611.0244
2019-06-11 15:09:38 [26830] loading URL table from "/var/squidGuard/blacklists/finance/banking/domains"

…

2019-06-11 16:32:29 [4165] database status: up to date
2019-06-11 16:32:29 [4165] license status: unknown
2019-06-11 16:32:29 [4165] configuration status: fatal error(s)
2019-06-11 16:32:29 [4165]  
2019-06-11 16:32:29 [4165] Changing daemon status to "error"
2019-06-11 16:32:29 [4165] time definitions are used; evaluating current ACLs
2019-06-11 16:32:29 [4165] A FATAL ERROR OCCURRED: all requests are answered with "OK" (see previous lines with "FATAL ERROR" for more details)  *****
2019-06-11 16:32:29 [4165] releasing the rwlock

i’ll try to replicate on a clean vm

You probably have a custom category without any domain (empty).
Despite the message logged this is not fatal. :slight_smile:
Please remove the last system call in /etc/e-smith/events/nethserver-squidguard-save/S2nethserver-squidguard-update-custom-list:

system('/usr/sbin/ufdbsignal -C "sighup ufdbguardd" >/dev/null 2>&1');

to fix the “double daemon” problem.

1 Like

Here’s the issue:

1 Like