If you would read the ones affecting boot and stuff - mostly corner cases and users not paying attention to prompts and of course there are bugs… But not show stopping. Which will be improved anyways. The story is - I trust the Debian community when releasing a major version. If you don’t, simply keep your desktop on Debian 7 till we hit the Andromeda galaxy.
I have switched to stable from testing on my AM5 desktop;
had a new fresh install on an ancient Lenovo notebook;
did an upgrade from 11 to 12 on a Macbook.
All worked fine and it is not like they are easy setups, sporting encryption, BTRFS, ZFS, extra firmware, games, lots of packages and so on. You do the testing for yourself if interested, if not don’t advise people not to.
So yeah, servers are next.
PS: how come every post of mine becomes a flame with you, unborn and a few others ?
This isn’t a flame. People generally don’t update working servers unless there is a need. Known bugs on new releases are generally for desktops where people are more willing to test.
Hope you read the now deleted posts. You also sounded dismissive and that is your prerogative. As is mine in return regarding your opinions - and yes, all the OSes become better in time, except CentOS that is, lol; so?
If that is the level of trust you have in your distro you should move to something else but in that case stop that bashing level of misrepresenting a good OS. You said you didn’t and wouldn’t even try it. What are you even talking about? Why post in the first place? Is your gratuitous opinion better than mine, using the packages since they were in the testing repo?
How are inconspicuous bugs are best discovered? In mass production I would say, otherwise never. Production by whom? Not you it seems. So please stop giving advises to people that don’t ask for them. The question was to the core team, more of a signal really, they responded and I am glad they did. So let’s move along.
Yeah I will just go ahead and upgrade some less critical servers maybe this week-end, than report it in Virtualmin forums so the team has some feedback and some starting points and so on. It is also a good indicator for the level of interest and usage. If not interested, please don’t recommend others to ignore it too.
I couldn’t even fully decipher what those meant. They were head scratchers.
I use Debian. It is my preferred OS. I have also spent hours at night trying to recover from server upgrades that failed. You’d be surprised by how many paying customers that can’t sleep if their services are down and seem to think calling and interrupting you every 5 minutes is going to speed up the recovery process. I have to guess you are running a personal server without the commercial pressures? No SLA’s to be concerned with?
This is why RedHat releases are what they are. Very bad for the desktop, great for server stability. That is why you see people on very old versions. They’ve probably been burned, as I have, on sever upgrades.
Ultimately though, the Webmin/Virtualmin team have to make sure their software doesn’t break anything for those that simply can’t afford down time.
Well stop explaining IT philosophies, again was not asking for them. And we are in the same industry aren’t we? Everything is in production, clients, friends and of course my personal playground, with the same kind of pressure.
So I would have the same problems. Except they are imaginary - never in 20+ years since I had a server was like: hey, updated Apache, now Apache is not working at all anymore, so I guess I will set myself on fire. It is not simply the OS but the whole ecosystem, and that’s why we end up favoring stuff. There were tons of problems sure but solvable because people are encountering them moving forward not staying still.
If you love Debian than you know: .1 releases is the moment to get into it or at least try to port stuff. Forgive me for drawing attention to one of the biggest and really the most courageous Debian releases ever, that will get a .1 release in maybe 1 month.
Rant: as for nowadays Red Hat I have only but words that would get me banned. For their model the same - all the new hardware I ever bought had problems due to their “keep it old” stability idea. So in fact was the other way around, that direction caused me only problems. 40Gb/s card - nope sir, we are still on 2.6 kernel! New CPU - hey, those sensors you need, yeah, they will not be there. And don’t get me started on the Let’s Encrypt bullshit Joking, but that is the spirit and I am so over it especially since the CentOS fiasco. And now they fired the Fedora manager, great support for Open Source in the RH camp.
Debian is the way to go for me so making it more like CentOS is a no no. But if you want to - sure, simply don’t upgrade.
I have been testing today on Debian 12, and fixed a few bugs in Virtualmin Config which now will enable Fail2ban work correctly.
Webmin version 2.021 supports SpamAssassin 4.0. So, we should be good to go!
I haven’t noticed (yet) anything else what isn’t working on Debian 12. So, if you’re interested in also trying, just use existing Virtualmin installation script. Note that you will need to manually fix (for now) [postfix-sasl] jail in /etc/fail2ban/jail.local file to look like this:
Installed VM on deb 12 yesterday in the office to see how it is looking, other then the fix Ilia posted above to get F2B working, the only other question I have is the system logs seem to now be missing (system log viewer is there). Is this something on the Virtualmin side or something new in Deb 12 ?
This is not a comprehensive review as its only been running for 5 or 6 hours.
I didn’t install Debian 12 yet, but this might be related to a switch to systemd and it’s binary logs. As far as I know, rsyslog should be still available on Debian 12 though … but maybe in a different way.
Anyway, I don’t know if this is on the roadmap, but maybe it’s now time to look into how Webmin handles system logs because it looks like retirement is near for rsyslog.
It’s already done. There is a new module System Log Viewer for this purpose, which supports reading/searching logs with journalctl as well as viewing text logs.
The journalctl support is still pretty rudimentary, and doesn’t include stuff like filtering by unit in a reasonable way (you can search by keyword, but that’s not the same), or stuff like specific date ranges.
Sorry for the late reply, I was in a coding spree.
If syslog is no longer installed by default in Debian 12, neither you or us should do anything (i.e. don’t try to install it). Use System Log Viewer instead. journalctl is the new source of truth for querying and displaying logs from journald on systemd systems.
huh? what? just because debian systemd people made it an optional package for bookworm, doesn’t mean rsyslog is in retirement…
on the contrary. rsyslog is alive and new releases come out very often…
sysadmins that prefer human readable logs, will stick with that, or syslog-ng or some other syslog daemon.
i usually go for Storage=None in journalctl, as it’s not very helpful service… it’s search options are often worst that a simple oneliner on regular log files…
anyway, personal opinion, but hopefully webmin will support both syslog and journald binary files. (?)
Oh … how could I have missed this? OK, super … sorry for this useless request.
I’m glad to read this! I admit that I don’t remember where I read that rsyslog was close to retirement in Debian. From memory, it was an exchange dealing with logs between two Debian experts… if I find the source, I’ll post it here.
At the moment you would need to use search to filter out the service you need.
Perhaps, in the future we could utilize -u option, e.g. journalctl -u postfix.service to specifically show logs for the give systemd unit.
Although, I don’t see the reason for doing it, as the output isn’t different between using -u option and filtering the results directly, i.e. using journalctl | grep -i postfix command.
?
Joking, but that is the spirit and I am so over it especially since the CentOS fiasco. And now they fired the Fedora manager, great support for Open Source in the RH camp.