| mason | Ooh. | 05:09 |
|---|---|---|
| mason | So, fail2ban needs to be forked. | 05:09 |
| mason | Has a bunch of systemd noted by default. | 05:10 |
| mason | Not sure that's what's biting me at the moment, but paths-arch.conf sure says to use journalctl for a bunch of stuff. | 05:10 |
| mason | Ah, that actually means arch the distribution. | 05:44 |
| mason | FWIW, https://codepoets.co.uk/2019/fail2ban-debian-buster-10-and-iptables-vs-nftables/ | 06:01 |
| mason | The correction noted fails on Beowulf for reasons I haven't sussed out yet. | 06:01 |
| mason | Ah, it's just that the behaviour has changed. Sigh. | 06:16 |
| Centurion_Dan | nft is a much better approach, but quite different from IPtables as far as the config syntax is concerned. | 06:18 |
| mason | fail2ban still defaults to iptables. I do want to learn nftables syntax, as it'll do things like let you speak to packet options, which can be useful. | 06:20 |
| mason | Used to do that with ipf and had to stop when I migrated to Debian from NetBSD. | 06:20 |
| mason | Rather, from NetBSD to Debian. | 06:20 |
| mason | Anyway, fail2ban doesn't need to be forked, lacks the bugs I thought I saw, and it all boils down to my wild confusion, as per normal. | 06:22 |
| mason | Or, maybe a change in fail2ban behaviour, but anyway. | 06:22 |
| onefang | Fail2ban often fails to ban, can be a struggle to get it to work correctly. | 10:25 |
| bgstack15 | The fix for lightdm in Ceres, which uses pam_elogind.so, has been requested in Beowulf. | 14:03 |
| bgstack15 | Beowulf version is at 1.26.0-1, and Ceres version is 1.26.0-7+devuan1. | 14:03 |
| bgstack15 | So the base version is the same, but Debian has had a number of package rebuilds. Should we add in the two fixes to a 1.26.0-1+devuan, or just CLOSE WONTFIX, or slide Beowulf up to the latest package revision? | 14:04 |
| LeePen | bgstack15: My suggestion is that we plan to include it in an early point release for beowulf. | 14:08 |
| LeePen | But I am reluctant to change lightdm in beowulf just before release. | 14:08 |
| LeePen | The manual fix is already documented in the Release Notes. | 14:09 |
| LeePen | But others may have different views? | 14:10 |
| mason | onefang: The funny bit is that it worked fine on Stretch, and its failure to recognize the stuff that pounds on my mail servers otherwise was evidently an intentional choice to work around some service someone uses that misbehaves. Amazing stuff. | 15:52 |
| mason | onefang: That said, I'm not sure any more that it's something we want to fix, as opposed to something Debian wants to fix. The Debian packaging of it is historically terribly. | 15:53 |
| mason | terrible, like my spelling | 15:53 |
| bgstack15 | Would anybody be interested in a freeipa-client maintainer for Devuan? When one true-holes the systemctl commands hardcoded in the freeipa client logic, it works just fine. | 20:48 |
| bgstack15 | I'm sure some additional work could be done to replace the systemctl commands with equivalent service commands. | 20:48 |
| amesser | Haha, I finally made DOS and Windows 10 talk to each other and have a network share :-) | 21:03 |
| fsmithred | Do they have anything in common to talk about? | 21:13 |
| xrogaan | still can allow any CON file | 21:54 |
| xrogaan | can't* | 21:54 |
| amesser | it is for an old sterilization unit at a hospital | 22:01 |
| amesser | its running dos and needs to pass the logs to the network for qm reasons | 22:02 |
| amesser | actually ist the facility not in my hands, it would use linux instead of win. however, it must be maintainable by the locals, and they only known windows :-( | 22:03 |
| bgstack15 | *and dos | 22:03 |
| amesser | I was surprised that intel has still dos drivers for gbit cards :-) | 22:04 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!