On 5/20/2021 2:17 PM, Anand Buddhdev wrote:
You could also log directly to files (bypassing syslog), and then have
some process follow the files and send the logs to a remote server.
This seems rather inefficient, but there are established and flexible
tools to do just this.
Without changi
Hello John,
I think it should be possible to use chroot and have there custom socket
mapped directly to rsyslog.
bind-chroot should be available in CentOS, try running
named-chroot.service instead of named.service.
I have not tried it on real installation, but I guess it should be
easiest way to
If you can have BIND log directly to a file, couldn't you use a FIFO
(prwxrwxrwx) or Unix domain socket (srwxrwxrwx) and avoid the disk I/O by
sending the log data directly to the forwarder? (E.g., Pulse Audio listens on a
socket for audio data from an application, and sends it in real-time to t
On 20/05/2021 23:34, John Thurston wrote:
Hi John,
> My subsequent read of the docs indicates that BIND on CentOS 7, while
> being told it is sending to 'syslogd', is sending to 'journald' which is
> handling all the messages and forwarding them on to 'syslogd'. I don't
> want journald handling m
Many years ago, when we ran ISC BIND on Solaris, we created a logging
channel to send the logged-queries to the local syslogd. We then had our
local syslogd forward most of the traffic on to a central syslog server.
I just tried to re-implement something like that on CentOS, and thought
I had
5 matches
Mail list logo