I’ve been with SuSE since before Caldera, then onwards.
I’m somewhat confused about the upcoming changes, although, I guess I can understand the need for SLE to compete with RHEL.
I’ve tried Cockpit on other platforms; If this is what replaces YAST I feel sad. YAST was more of a ‘SMIT (R)’. Cockpit feels like an unfinished package, written by SUN or Veritas. If you can figure out it’s hierarchy and you have twice the time to spend navigating its labyrintheen architecture, and then another chunk of time using the CLI to configure what it doesn’t, then it works. I’ll miss YAST
Cockpit, as said, I can do without.
Zypper, there may be a reason to change, but I am not constrained by sequential update. It works and I know how to use it, and it is plenty fast. Why change?
SYSTEMD - going forward is going to be a block for me. I am running Linux, a Unix like OS. If I wanted to run SYSTEMD-OS, I would.
Now, you may ask, does it matter what I think? Probably not. But #4 above is the killer for me. SYSTEMD is made in the image of Blue; it offers 10% good functionality, but 90% bloat, and lockin. Please SuSE, consider.
TasP
@70tas systemd is here to stay… you didn’t mention the move to SELinux and systemd-boot…?
What do you need to do with YaST that is missing in Cockpit?
To be honest once my systems are setup and that is generally done via a script since I have multiple systems with differing hardware and uses. It’s just generic maintenance that might be needed, but very seldom use YaST…
There is no way back. systemd is the per se standard since a decade and Sysv init is deprecated since ages. There are no advantages of keeping deprecated stuff in a modern operating system. systemd is faster (way better boot performance), easy to configure services, provides logging and much more.
You should do a small research and will notice that all major distributions use systemd (since ages).
Not sure what you tested, but Cockpit on openSUSE is straight forward. A leftside menu which opens the content on the right side when clicked…
I disagree that it is faster. For SuSE there may be no way back, but consider the ramifications of the new changes. Win…s, sys…d, one and the same. As for licensing they are coming to parity.
Since ages? don’t know about ages, and it does mean that they can more easily create a distribution, since systemd will take on more and more of the load. If that is what you think an OS should be, then all will be well with the world. I’m concerned that he who bunks with the devil, well… I see a day when there will be two OS’s, Win…s and Sys…d. That’s all I’m saying. This is my opinion of course, and everyone is entitled to theirs, and I respect that.
I didn’t say all systemd was bad, I said there is some good, but a lot of lockin. As for SELinux, that may be a more adequate system than the alternatives. But I can’t say I am a security expert, so I leave that to those experts.
I don’t know why you feel there must be conspiracies… I certainly didn’t say that Hui. I don’t think there are conspiracies, but I am from the old school, where a tool should do one thing well. I started with an old Unix variant of GCOS, don’t remember the version.
Smaller chunks of code, that is how Open Source or PubDom can verify and validate code. Not that SystemD has malicious code, but the larger the code base the easier it is for bugs to creep in. I think we can all agree on that.
And guys, I didn’t intend for an argument, just wanted to express my opinion and leave it at that. You are free to disagree with them, of course.
And thank you for the replies.
As for the change to systemd…the RFC for that change started in December of 2011 on the factory mailing list. It’s been part of the distribution, as hui said, for over a decade.
I think the period for comments is long since closed. systemd is here to stay until something better comes along.
The horse has left the barn, and sysvinit is probably not coming back.
YaST, as has been noted in several other threads here, has become difficult to maintain with the recent ruby updates. Many modules are no longer usable even today, apparently. If you don’t want it to go away, you need to either help contribute to its maintenance or help find people who are interested in continuing to maintain it.
If it doesn’t get maintained, it doesn’t get included, and it seems that those who were working on it are now putting their time and effort into other things (including Cockpit). So you can either work to either help maintain YaST or find people who are willing to do so, or you can provide feedback as to what features are missing from cockpit so they can be tracked and considered for addition.
Things change in Linux distributions, and change is always hard. Familiar tools get replaced with unfamiliar tools, and it takes time to adjust.
As for zypper…I’ve not seen any indications that zypper is going away. Maybe I missed it, but there is a fair amount of active development to improve zypper (parallel downloads and single transaction installations, to name two huge improvements).
Almost everything lol. But I’ll tender a small list:
Console management - Things like workstations can almost all be managed with UYUNI, but a ton of simple changes like printers are way easier with yast and will certainly be a breaking point for me when YAST is fully depreciated.
Printer Management
System Configuration editor
Bootloader
Enable/Disable VNC Remote Administration
Installation of Hypervisor tooling in an easy way (Honestly, I can probably figure this out in 5 minutes with a zypper pattern, but its still a learned habbit)
About 20 other weird tiny things that annoy me greatly, but yast makes better.
Honestly if cockpit gets some sort of TUI interface I’ll probably do fine, and if YAST doesn’t get fully removed until 17 so cockpit can develop more (I’ve used it for years and the change from a few years ago to now is awesome, but still needs more plugins).
As for the sytemd, I love it/I don’t care at all. As long as I can get my job done it doesn’t matter to me, and openSUSE/SUSE products have always made it easy to get my job done.
Anyone still moaning about systemd, should be sentenced to writing init scripts for everything for at least a year
Betting on YaST future maintenance is betting on a dead horse. Just checked, I don’t have it installed since I removed it ~1 year ago and have not missed it one single moment.
What is this phenomenon where the same people want the latest and greatest, yet want to be stuck with outdated Frankenstein source code?
And all said and done here: No one is forcing anyone to use a linux distro, there’s still some niche distros with sysVinit, mkinitrd, maybe even with cu ( old Call Unix ) and so on. Oh wait, that doesn’t have GNOME4x, Plasma > 4, eh …
People have fond memories of “the good old days”. But they mostly remember the good parts and forget the bad parts. Because of that selective memory, I tend to describe it as “the good old days that never were.”
@oranclay I’ve not used YaST for a printer in eons, now I use a HP LaserJet, hplip is there (well hopefully a QT6 version…) still have to install the plugins via hp-plugins gui or I prefer cli and localhost:631 to manage…
Not a lot to touch there these days? Perhaps some profiling for say tuned, but that can be done via Cockpit.
Hmm, well again for me that’s all scripted to add systemd-boot options, way easier than having to edit a grub file…
That is integrated into the GNOME desktop, so no need for YaST?
Yes, again a script option for me rather than installing a pattern, but Cockpit integration is there for virtual machines, podman etc.
Transactional update (cockpit-tukit) is available in Cockpit?
I know a lot is available for some web management, the issue I always roll back to in my environments, is that it is 100x easier to ssh into a clients machine, edit whats necessary in yast and move on with my day then to rely on other more automated options.
The automation is great for setup and general configuration (I basically live on UYUNI now days), but web management options are just a second rate configuration option for workstation management.
An additional note:
For the level of problems made easier in TUI, it would probably be easier to incorporate them into jeos-config given how small the list is becoming. That’s why I didn’t list things like network configuration and 389DS management. They are broadly getting good enough in browser, and we have basic tui tools to solve normal network configuration issues.
When I was new to Linux YaST was possibly the first reason why I chose openSUSE. When I had to switch to Ubuntu for HW compatibility reasons I greatly missed it and switched back to openSUSE as soon as my HW was adequately supported.
Now, some 20 yrs later, I can likely live without it but I think that a newcomer still faces a steep barrier without System Configuration Editor or Hypervisor Installer. Cockpit looks good for sysadmins but doesn’t look as user friendly to a newcomer.
Anyway we might not have as many newcomers anymore, they are just buying a phone and use whatever BigG throws at them, not even bothering about configuration and such
Yeah, I am usually harsher: “The good old days where we all had mumps, measles, proper poverty, starvation, ~1500 people/year dying of food poisoning, no diversity, proper legal racism”. And so on.
That has traditionally been my argument as well, but that is also why I’ve spent the last year validating system designs for a business openSUSE layout (and even successfully got a whole company running only openSUSE).
While I love YAST for exactly what you described, I’ve also found that the complete lack of direction like “Here is a method to get everything running” is more of an obstacle than the system configurations them selves. I think the remaining tools outside of that basic design model could easily be handed off to jeos-config and not have too much loss of function.
That’s a subjective comment. I see nothing that makes YaST particularly any more friendly than Cockpit - maybe more familiar to those already used to it.
Many user config are not in YaST’s domain anyway - often just desktop related, and handled there (eg System Settings). As Cockpit is being actively developed (larger user base), bug fixes and improvements are far more likely.
They still have all of that.