On Wed, 4 Dec 2019 21:46:23 +0100 Andrea Borgia <and...@borgia.bo.it> wrote:
> Il 02/12/19 01:32, Charles Curley ha scritto: > > > > How do you know this? Granted, upgrades are a likely culprit, but > > don't > > I log all config changes, timestamped and with relevant links to > documentation or supporting evidence for the decision. OK, sounds good. > > > > Is there any reason you are doing production backups on Debian > > testing? If these are production backups. > > mononoke is my home desktop, clarisse is the laptop and flyingraspi > is a raspberry running an ADS-B monitor, so hardly "production". > Still, it's annoying when backups fail. Ah. Your choice. I consider both my laptop and my desktop to be "production", as I use them both daily. Perhaps I am over-cautious. > > > > If I had to run one of the two, I'd prefer xinetd. However, I > > suggest you use SSH for all your machines. It's tedious to set up, > > but once set up it is stable and more secure. Then you might > > (depending on what other services it manages) be able to take > > xinetd off entirely. amanda-common requires inetd, so you can't > > take it off. But you can use systemctl to stop it and disable it. > > I have verified that the xinetd config was a left-over and removed it. OK. So we are working with inetd. > > clarisse is actually using the ssh connection precisely because it's > more exposed than the other two hosts and I didn't want to punch a > hole in the firewall for such a limited scenario. > > The decision to leave the other two with bsdtcp connections was based > on raw speed. Fair enough. -- Does anybody read signatures any more? https://charlescurley.com https://charlescurley.com/blog/