ConfigServer Community Forum

When using the cluster function csf --ctempdeny the comment is missing from the entry that is added:

To test:

csf --ctempdeny 203.0.113.1 3600 -p 3306 -d in 'This is the comment'

The entry that is added across the cluster:

lfd - Cluster member XXX.XXX.XXX.XXX (CA/Canada/web0.myserver.com) said, TEMPDENY 203.0.113.1, Reason:

I tired to enable the new docker support after upgrading to CSF 12.00 and remove our old csfpost.sh but I get this when I try to restart CSF on our Plesk server (using csf generic). It looks like the docker0 device isn't being passed to the new rules? In our upgraded csf.conf there was no DOCKER_DEVICE setting present in the config file.

# csf -r
Bad argument `MASQUERADE'
Try `iptables -h' or...

Hey chirpy,

It looks like there is a javascript error on the cloudflare UI page on CSF v11.01 which prevents the UI from working properly. In Firefox and chrome I see this in the browser console on the cloudflare UI page:

TypeError: $(...).chosen is not a function yMq3wnv0pfUc5kjFZIOxejZkGj:44:3

Once javascript breaks the code below doesn't get a chance to re-enable all buttons even though...

In the GUI (using WHM) for CSF, there is a simple wording error for the English language description.

Option: RECAPTCHA_NAT
If the server uses NAT then resolving the hostname to hosted IPs will likely not succeed. In that case, the external IP addresses must be listed as comma separated comma separated list here

Specifically: as comma separated comma separated

Thanks!

lfd spawns a secondary process to handle &integrity()

This secondary process, when it encounters an error (in this case, inability to send a report because it can't reach the configured smtp server), calls &cleanup which subsequently calls unlink on the pid file. This then causes the pid file check near the top of the main while(1) loop to fail, which causes lfd's main process to exit.

In this...

After the 12.05 update my lfd log is full of exim syntax errors:

Aug 10 08:17:20 vpc5 lfd : Exim syntax errors from

resulting in my ingress mail servers to be blocked.

I see this in the changelog:

Added new regex for LF_EXIMSYNTAX

I whitelisted my ingress servers to get mail flowing again but any help to fix the issue would be appreciated.

I received these errors in /var/log/lfd.log when trying to use an Include in csf.dyndns

Apr 23 16:03:41 mail1 lfd : DynDNS: Lookup for failed
Apr 23 16:13:33 mail1 lfd : DynDNS: Lookup for failed
Apr 23 16:23:34 mail1 lfd : DynDNS: Lookup for failed
Apr 23 16:33:34 mail1 lfd : DynDNS: Lookup for failed
Apr 23 16:40:13 mail1 lfd : DynDNS: Lookup for failed

Commenting out the 'Include'...

I attempted to use an include in csf.redirect and received this error on executing csf -ra.

Error: csf: Incorrect csf.redirect setting ([]): , at line 3172

I removed the Include line and started csf with no issue.

Point of clarification: The readme.txt specifically states not to use comments at the end of the lines in the include file but mentions nothing about comments on lines by...

I am getting this error when restarting/reloading the CSF firewall:

csf -r
*WARNING* ip6tables nat table not present - disabling DOCKER for IPv6

This happends on 2 servers with Debian version 8. No such errors on servers with older versions of Debian.

I've just enabled MESSENGERV2 on two servers and tested it, I get this php error: Warning: file_put_contents(/var/log/lfd_messenger.log) : failed to open stream: Permission denied in /home/csf/public_html/index.php on line 69

File permissions:
-rw------- 1 csf csf 0 Feb 18 03:51 /var/log/lfd_messenger.log

I think the problem is that php runs under the user of the domain on which the error pops...

It seems like there is some issues with MESSENGERV2 using the wrong SSL to show the recapcha.

Thank you.

It looks like there was a change in the cPanel ACLs of 'software' to 'thirdparty' - - going from version 66 to 68. The ACL that CSF tries to set is currently 'software-ConfigServer-csf', but the 'software' ACL is no longer available.

This has broken reseller access to the CSF UI in WHM.

If any additional information is needed, please let me know.

The mod_security v2 (apache) regex (line 335 of RegexMain.pm) has a bug that prevents it from working with IPv6 addresses.

The rule matches the first block in the log entry, which contains the source port appended to the client IP. The rule does attempt to remove this source port, but only does so if splitting the IP by the ':' character returns a list of size 2. Obviously, this won't work for...

Hi,

it is stipulated in readme.txt:

27. CloudFlare
6. URLGET must be set to 1 (i.e. LWP) must be used

the problem is that 1 is HTTP::Tiny and 2 is LWP::UserAgent

so what is the correct setting?

thanks

I think that the readme.txt is missing some info at the end?

7. To add a temporary block that blocks in csf as well as CloudFlare:

Then it ends.

Hello all,

I tried Fedora for the first time today, and CSF failed to install correctly (I've been successful with Centos for quite some time).

Basically the systemd unit files, and the binaries get placed with incorrect SELinux context. I fixed the contexts in /usr/lib/systemd/system/ before I could copy them, but in /usr/sbin they were erroneously:

unconfined_u:object_r:user_home_t:s0

The...

With MESSENGERV2 and MESSENGER_HTTPS enabled, any time a SSL certificate path changes (which with AutoSSL could be any time the AutoSSL process runs), the csf.messenger.conf file is then pointing to a non-existant file, which causes Apache to throw a syntax error and not start.

Unfortunately, it seems like there isn't currently a cPanel hook that will always catch every instance of an SSL...

I noticed that the new lfd_messenger.log has messenger spelled as Child Messanger :

Jul 29 00:00:03 www0 lfd : Child Messanger (HTTPS): TERM
Jul 29 00:00:03 www0 lfd : Child Messanger (TEXT): TERM
Jul 29 00:00:03 www0 lfd : Child Messanger (HTML): TERM
Jul 30 00:00:06 www0 lfd : Child Messanger (HTTPS): TERM

We noticed random servers in our csf cluster were not responding when sent DENY commands.

Restarting at cli with csf -ra reopens the port:

[ root@cp1 ~># csf -p|grep 7777
7777/tcp -/- - (9359/root) lfd Cluster Server /usr/local/cpanel/3rdparty/perl/524/bin/perl

But just a few minutes later, it closes again:

[ root@cp1 ~># csf -p|grep 7777

[ root@cp1 ~>#

There doesn't seem to be any...

We have device names with a dash in them:
# ifconfig
eth-internet: flags=4163 mtu 1500
-- SNIP --

eth-local: flags=4163 mtu 1500
-- SNIP --

lo: flags=73 mtu 65536
-- SNIP --

eth-local is a management interface and I want it to be skipped by csf, so I set ETH_DEVICE_SKIP to eth-local, however during start csf reports that this is not valid:
*WARNING* ETH_DEVICE_SKIP device not listed in...