Hacker News new | past | comments | ask | show | jobs | submit login
Suse is once again an independent company (techcrunch.com)
364 points by bsg75 on Mar 17, 2019 | hide | past | web | favorite | 150 comments

Being owned by a Private Equity fund can really not be described as being "independent". Such funds have a typical investment horizon of 5-7 years, with potential exits being an IPO, a strategic sale (to a bigger company) or a sale to another PE fund, with the strategic sale probably more typical. In the mean time the fund will impose strict growth targets and strong cost cuts.

Im working for a company that was bought by a private equity in 2015 and taken off the stock market. Then introduced to the market again about two years later.

They left us mostly alone, and most things actually turned to the better. Maybe not having to answer to thousands of shareholders gave management the possibility to implement some needed changes. One example was a new salary/bonus/retirement structure giving a huge boost to most of us. So YMMV.

good point - there is far less reporting and PE has a much longer time frame than the 2 quarters everyone but Warren Buffet cares about

Yeah, I'm not sure how anyone can call private equity "independent". Our whole last year had selling the company as our top priority. Not something I'd choose in a truly independent position.

Being independent doesn't mean "without owners". That the owner is PE is actually quite a good position to be in, as others have already commented. So yes, "independent" is the right word here.

I think it really depends on a couple of important variables, including cash flows and company management. Our private equity overlords care mostly about the macro goals that are generally well defined in advance (sales targets, M&A budgets, bonus pool) and our CEO has done this a few times so he has a track record of predictable returns, which is primarily what a PE fund wants, vs the hockey stick growth of VC money.

In this economic climate and especially outside of the US, PE gives you a lot of the benefits of going public without the extra scrutiny and reporting. I've seen several established companies in my industry forgo going public and unlock liquidity via private equity or sovereign wealth funds.

I can't be the only one surprised with the $2.5 billion acquisition?

Suse basically fell off my radar a decade ago, I thought the next announcement I was going to hear about them was going bankrupt and being sold off for parts. Guess they've been super quiet achievers and have managed to accrue an Annual Revenue of 300M that I've somehow missed behind the release/announcements of flashier distros, but here we are with news of a 2.5B acquisition - congratz to the team.

Indeed, SuSE has a long history of substantial engineering investments: https://lwn.net/Articles/613006/

Based on this expertise, the only real competitor in their business is Red Hat.

For some weird reason Suse sees a lot of use in European countries with less English influence, but I haven’t seen any exact numbers in years.

In my country (Denmark), most enterprise runs Redhat, but my personal opinion is that Suse would make more sense after the IBM purchase, especially for public sector usage.

It’s not weird, they just have a decent sales muscle in their original markets. It’s actually weird how the US media insisted ignoring it even after the Novell acquisition.

SuSE originally was a German company, so it's not surprising that it got a foothold in German-speaking countries, being able to provide support and contracts in German etc. Years ago (late 90s or so), you could even buy boxed SuSE Linux in electronics stores in Germany, next to the Windows CDs.

I bought a boxed copy of SuSE from a store (Best Buy?) in the US in the early 00’s. I believe it had a green gecko on the front of the box.

They had a different color box for each version. Suse actually would mail you multiple boxes for free at one point!

Boxed copies were sold on other countries too, I still have one in my shelf that I bought in Norway in I guess 1999 or 2000.

As an employee of IBM Denmark I'm interested in hearing what makes you think so from a danish perspective.

IBM is my first employer and while my inside impression has been great, I'm wondering what the danish view from outside is!

In my experience IBM is one of the worst companies to do business with in Denmark, if your requirements and needs end up being different from what was specified in the original contract.

That’s not what I meant though, IBM is an American company, that holds inherent problems for the European public sector. Basically I think we should bet on our own, because recent years have shown us that we may not always have a great political relationship with the US. I think it’s highly unlikely that we’ll have a bad relationship with the US, but it’s a risk assessment we need to think about when we enter into multi-million or billion dollar decade long contracts.

> recent years have shown us that we may not always have a great political relationship with the US

IBM happily helped Hitler plan the Holocaust during World War II. Not even war could stop them from making a buck, so I doubt Trump's shenanigans would.

I haven’t used SuSe since version 4 or 5 I think.. do you see it as a viable replacement for CentOS?

It is, or rather it would be, if more cloud providers etc. bothered to provide images.

Indeed, I think that RedHat & SuSe are very similar in their ideas and in their usage patterns. Both contribute quite a bit to upstream projects (there is a lot of work being done on CNCF-stuff in both places, both are RPM-based and both offer a number of different "tracks" depending on your needs). openSUSE also has been very good at keeping up with things that are in development.

On the desktop-side, I think openSUSE is actually nicer, since they officially support a myriad of DEs (notably both Gnome and KDE).

For what it's worth, I use (and have been) using openSUSE both on my work laptop and my private laptop as my daily driver, and I use openSUSE on various private projects as a server OS, though I'm using CentOS 7 at work for all our servers because some of the software we use only supports Ubuntu LTS and CentOS (because the vendor is incapable of building non-idiotic RPMs/DEBs...).

> they officially support a myriad of DEs

For years, Suse was the de-facto reference implementation of KDE, where it was the primary desktop. I believe this changed after the Novell acquisition (but it’s been so long, I might misremember), and KDE had to find other ways to get an equivalent “showcase” distro.

Novell changed it because they had also bought Ximian, who at the time was one of the bigger GNOME development companies.

Unfortunately, that legacy remains with us today, and SUSE is not very friendly to KDE for the enterprise, which I think is really a shame.

First you need to distinguish paid vs. non-paid variants.

Suse has SLES and IBM has RHEL. I am not in the position to compare them.

The non-paid variants are OpenSUSE vs. CentOS. One big difference there is the support life. CentOS 6 with an ancient 2.6.32 kernel and upstart came out 2011 and is still supported. OpenSUSE releases are supported ~ 1.5 years. There are different opinions whether upgrading from one OpenSUSE release to the next one should be done or not. I typically prefer new installations because they result in cleaner systems. But I also upgraded once an it was painless.

CentOS is RHEL, just without the branding. SLES and openSUSE share the same core packages. So you get the stable, well tested enterprise stuff.

Correct. I'd say openSUSE is more akin to Fedora in that context as both base and development distribution for SLES, much like Fedora is for RHEL (which in turn then "becomes" CentOS).

There are 2 OpenSUSE variants, LEAP and Tumbleweed. Their release model has changed maybe 3 years ago.

LEAP is exactly the same package versions as in the newest SLES I believe. I would expect them often to be a bit older package versions than Fedora, but I haven't compared recently.

Tumbleweed is a leading/bleeding edge rolling release distro. The closest equivalent is Fedora Rawhide.

Sure! They have the same goals and both have a good track record. CentOS has the market share advantage.

the SAP team at my old workplace had a bunch of SUSE merch sitting around. Suspect this might be a data point.

A couple of years ago (last time I worked with it) the SAP developer server images for consultants (iirc. it was preconfigured ISOs, it was a bit odd) ran on SUSE so there's a tight connection between the two. Since both have German origins it's not surprising.

Indeed, SAP wants you to run their products primarily on SUSE Linux Enterprise Server, with RHEL being an "also ran" (despite enjoying full support from SAP).

This despite SLES' controversial and never-working-quite-right "vm.pagecache_limit_mb" kernel parameter (which they're finally ditching in SLES15, I hear).

Don't get me wrong, SLES these days is a mature, high-quality distro (I'm working with hundreds of installations running SAP products), but sometimes they do weird things, like bumping kernel version from 2.6 to 3.0 within one service pack (sometimes back in SLES10 or SLES11 days, iirc).

Linus talks about the 2.6 -> 3.0 change here: http://lkml.iu.edu/hypermail/linux/kernel/1107.2/01843.html

>As already mentioned several times, there are no special landmark features or incompatibilities related to the version number change, it's simply a way to drop an inconvenient numbering system in honor of twenty years of Linux.

All they said was that it was weird. Even with the explanation I'd still agree that it's weird.

"a bit odd" seems normal for SAP (where some value of "a bit" is "a lot")

Yes, SAP seems to be the main selling point ... "SUSE Linux Enterprise Server was the first OS for SAP HANA and SUSE continues to be a co-innovation partner for solutions such as SAP S/4HANA, SAP Cloud Platform, SAP Data Hub and many more. ... "

... and the cloud versions of these products are SuSE based as well.

Original announcement: https://www.suse.com/c/news/suse-partners-with-growth-invest...

Congratulations, SuSe!

Edit: The title seems to be a little mischaracterizing, SUSE themselves call it a "Sale to EQT". But I suppose there is no other possibility.

It’s now owned by a private equity company. So their expectation is that it will grow and make money and they control the board, but the PE company itself doesn’t have a technology business that will determine Suse’s path.

If it were a tech company buying them, the assumption would be that the buyer would try to digest Suse and incorporate its components into the parent.

Every independent company has an owner or owners. The point is that it has been split off once again from the rest of Micro Focus.

openSUSE is a really good and modern linux distro, people complaining about here probably haven't used it in a long while. Whilst i worked in A&T we handled a few thousand+ servers running SLES 11 SP2 and it was solid.

I think they're biggest handicap is they're not from the US and while i worked there people acknowledge they have horrible marketing. Also being bought and sold a few times doesn't help, and with the HPE acquisition a few years back, a lot of employees (on HPE side) simply left.

The kubic project has developed the only viable alternative to CoreOS that's being phased out. Unfortunately where MicroOS falls short is SUSE's stubbornness for using BtrFS, which causes a lot of headaches and has been dropped from all major distros at this point. People in the kubic project have also been contributors to kubernetes, kubeadm, OCI, (they even contributed a lot of code to podman, from Red Hat) and they've been a significant driving force for rootless containers https://rootlesscontaine.rs/. It's weird they're not more often in the limelight with all the hard work they do, but maybe they'll take this opportunity to fix their marketing and steer into a better direction :).

The problem with kubic was the unknown space it was in - was it an appliance, or a generic distro with kubernetes packages. The btrfs stuff could have been quite good if they shipped btrfs snapshots as the update mechanism (like coreOS), but there was the pressure from the SLES/openSUSE side to let deployers install random software as needed, which meant you were left with the mess of "zyyper dup" + snapshot + reboot, and not having a known version for doing config updates.

The team behind is really smart, and doing good work - it just needed a forceful vision imposed on it for it to go somewhere.

btfrs is such a weird choice for an enterprise focused system where people mostly want reliability / set and forget, though maybe the bad rep is old now (I haven't used btfrs in years).

It is used for the root filesystem, which gives you nice functionality, like the package manager or YaST automatically taking snapshots in between system changes, which means you can then audit changes and rollback.

It even allows you to boot the system into one of these previous snapshots.

an engineering "teacher of teachers" I spoke with was able to corrupt btrfs with a few well-chosen strokes, as it was told. The implication was - not ready. This was three or four years ago.

I'm still reluctant to trust it, but there has been a lot of work on btrfs in that time frame, particularly around reliability and durability. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...

What a coincidence. Just the other day I was going to take a trip down memory lane and fire up a few VMs of my first Linux distributions. Then I realized Suse 7.3 (I had a physical box set) wasn’t freely available in ISO format anywhere. Bummer.

Still going to hunt down an early Slackware ISO and might try the first versions of Fedora and Ubuntu again.

Anyone remember Yggdrasil Linux?

Yes, being sold via the Walnut Creek CD-ROMs.

Soft Landing Linux on a million floppy disks?

Boy are you guys sending me down memory lane. I remember all of these.

The version I tried was bundled with the "Linux Bible" book

It was my first exposure to linux; the 1994 version.

Here you go:


The ISOs on that mirror go back to 12.0 but you can roll your own using the older releases.

I think Suse came from Slackware originally.

Yes. IIRC the very first version was basically a german translation of Slackware.

Yes, SUSE indeed started as a modified version of Slackware. For a long time, SUSE's package categories names were based on Slackware's disk sets names (even after rebasing away from Slackware). Later, SUSE hired the author of the Jurix distribution and rebased on top of that. SUSE and Jurix were both from Germany.

History of Jurix: http://linux2.mathematik.tu-darmstadt.de/pub/linux/distribut...

SUSE had a stronghold in the nineties in Europe. E.g. here in The Netherlands, many bookshops would cary SUSE Linux boxes.

In Norway as well, at least in academia - I remember seeing SUSE boxes in the bookstores at both the Norwegian University of Technology and Science and the University of Bergen - this would have been in 1996-1998 or so.

I thought it was spawned off Red Hat. What is the story behind Suse's use of RPM?

Redhat open sourced rpm. SuSE adopted it because they saw the benefits.

Try 6.1 noob :)

try 5.3 . The old good YaST rulez

Yast was never good. I still have nightmares of it happily overwriting and destroying my configuration on several occasions. No warnings shown or questions asked.

Oh, memories! Does Suse still use yast or did they replace it?

openSUSE still "uses" YaST, as in, if you choose to change something via YaST, it will happily do so. You can also simply install/uninstall various components. It also happily integrates with changes made elsewhere (e.g. the "Firewall"-module wraps around firewalld nowadays, which you can influence either via YaST or firewall-cmd etc.).

I sometimes use the Package Management part of it, since searching for something is sometimes quite nice in it. That said, I don't use it for installing software or doing updates etc., so you can absolutely use e.g. zypper & YaST side-by-side.

Oh, it's still there, we just never use it. Or mention it in polite company.

I am personally quite happy for this, I found SUSE in 2002 as it was installed on the computer I got at that workplace. I have never really left it since. Zypper, yast and Open Build Service are strong points, and lately I have used the Kiwi image creation tool. Solid too. I hope the best for this distribution.

This is good news. With IBMs acquisition of Red Hat it's valuable to have one of their direct competitors be fully independent.

Canonical is also fully independent..

Personally I've been on Ubuntu for over a decade. But the Telecom industry that I work in is entirely dependent on RHEL and SLES. A few large players (Like my employer) maintain their own distros based on CentOS but I am yet to see anyone use Ubuntu. From what I've heard, it's the same with the Enterprisey industries. So, I'm not sure if considering Canonical a direct competetior to Red Hat is accurate.

Just because it doesn't have market share doesn't mean it's not a direct competitor. It's like saying Apple isn't a competitor to Window since it only has 10% market share.

"entirely dependent" is much too strong. Many of the company's largest clients are in the telecom industry (source: on staff). Canonical is very much a competitor to Red Hat.

Canonical will be acquired by Microsoft. It's only a matter of time

I don't understand why Microsoft would acquire Canonical. Microsoft already has an operating systems business in Windows. I don't see any obvious synergies with other lines of business except Azure.

Azure officially supports a number of distributions. The risk of alienating the existing partners who package these distributions for Azure likely outweighs the small benefit Microsoft would get from acquiring Canonical. A deep partnership already exists (e.g. Ubuntu Advantage integration throughout Azure). They don't need to own Canonical for that.

> Microsoft already has an operating systems business in Windows.

Windows the OS is gonna die sooner or later, it's a question of time. Windows the legacy application API layer is probably gonna stick around forever.

Linux kernel + proprietary Windows application compatibility secret sauce seems like a killer combination. (Think Wine, except easy to use and actually working, and commercial.)

I think the only thing stopping Microsoft from following this plan is legacy devs and managers having a vested interest to keep their jobs around.

The Windows API may be old and rotting, but NT is a very modern kernel compared to Linux. It would make as much sense for them to support Linux binaries on Windows. Oh wait, they do ;). I think Microsoft would make quite a splash if they open source the NT kernel. They could even follow Apple's model, where they open up most of the kernel and keep win32 et al. proprietary.

I don't think Windows will die anytime soon. Yes, the traditional Windows desktop will die eventually. This is why Microsoft probably makes such a strong play with Azure and Office 365. But Microsoft can just continue to sell Windows + Blink-based Edge as a modern desktop for web apps that also has support for legacy win32 applications for those companies and individuals that need it.

Building a compat layer on top of Linux seems to make little sense. It will be a huge time investment and while desktop Linux is great for developers, the whole ecosystem is to volatile for most end users. Plus the Linux graphics stack is still not where it needs to be.

Whether they'll acquire a Linux distributor is a completely separate question for me. I think it makes sense to acquire know-how and some influence for Linux on Azure. On the other hand, maybe they don't want to upset other Linux vendors. Oracle, Red Hat, and SUSE are typically used on expensive SAP/Oracle deployments. In fact, they were even SUSE resellers one day ;) [1].

[1] https://www.computerworld.com/article/2532632/microsoft-to-b...

> The Windows API may be old and rotting, but NT is a very modern kernel compared to Linux.

I've seen this parroted before, no evidence to support this opinion.

If you're going to keep the Windows API then why would you replace the NT kernel from under it when it is, in many ways, better than the Linux kernel? What's more likely is that Microsoft buys Canonical so they can initiate the extend part of the EEE strategy with Subsystem for Linux.

Canonical would be a target mainly because of most - if not all cloud products - primarily run on Ubuntu.

Cloud Native products literally do not care for RHEL/SuSE and the like for a rather large portion but just give you Ubuntu. Also, paid support is given in combo with Ubuntu. Paid support: big business.

Can I get the source? MS and Canonical have shown strong partnership lately but that doesn't mean MS is going to buy Canonical.

There's no source for such a comment, it's speculation. The whole "MS will acquire Canonical" rumour has been floating around for at least the past 5 years.

The desktop side is problematic. Developing desktop Linux makes no sense for Microsoft, but acquiring Canonical and killing that part would look bad.

There isn't much left to kill there, given the massive reorg/layoffs in 2017; it's all outsourced to the community already.

It makes sense in that having more involvement in desktop Ubuntu development would make it easier to understand how to make WSL better on Windows.

However developing linux applications is still painful on windowd. Even with wsl the Fs perf bites pretty hard.

Still don't know how to pronounce it.

Age old argument, but it's pronounced like soo-suh. Reference: https://www.youtube.com/watch?v=M9bq_alk-sw (on the SUSE youtube channel)

Can’t Stop the Seuss-uh [0] demonstrates the pronunciation. They’re a German company.

[0] https://youtube.com/watch?v=A-Rn0iQEpc8

That's a seriously well made parody.

There’s a whole bunch of them on their YouTube channel. All very well made.

This one is great as well https://youtu.be/b0tsZB_LEQk

How about that one? Not SUSE related. https://www.youtube.com/watch?v=SYRlTISvjww

It's about SUSE kgraft, kernel live patching.


Just press the speaker button on the left-hand side.

I believe like the doctor.


Ess U Ess Ee

How are you pronouncing U?

Like in Suzie?

What about queen?

It was an almost legitimate question.

More like (Konrad) Zuse ;)

In most enterprises it’s pronounced “rehl”.

Haha, a comedian

SuSE Linux Enterprise Server is by far the most mature of all GNU/Linux distributions. Their AutoYaST provisioning system with Kiwi imaging capability is conceptually identical to Solaris' JumpStart™️ with compressed Flash™️ archives and is technologically superior to Docker images because it is much more simpler, yet offers the same conceptual capability without the need for cgroups.

Their technical support is wonderful and they truly deserve to be the most deployed GNU/Linux distribution in the enterprise. I hope they succeed in stamping out redhat because they have a superior product with far more competent engineering and support.

Wonder how this is going to affect Micro Focus’ other products in the future, they have a lot of software inherited from Novell that relies on SUSE (some of which aren’t even distributed as RPM’s but SLES-based virtual appliances like ZenWorks).

It's probably not going to affect it very much. SUSE has a decent sized chunk of appliance/embedded Linux business.

Aside from SAP, another user that most people wouldn't be too aware of was VMware. Before they created Photon OS, they used SUSE Linux Enterprise for their vCenter appliances. They've been long-time SUSE customers.

In addition to that, a large chunk of point-of-sale systems are built using SUSE Linux Enterprise customized JeOS ("Just enough Operating System") appliance systems.

I think this is likely to be just fine.

(Note, not a SUSE employee, this is just based on my spelunking and research.. ;) )

Also, if you work at a larger company and have a Siemens/Unify IP phone on your desk, it's not unlikely that it's connected to OpenScape Voice running on SUSE. (Even if it's not an IP phone, it might be connected to a smaller sub-PBX which itself talks to an OpenScape Voice server.

Loved SuSE forever..well at least since 1998.

I wonder if EQT's investment is interest in potential growth from Enterprise customers that may seek alternatives to IBM if their RHEL relationships change? Suse being the "other RPM distro".

EQT buying SUSE was announced before IBM buying Red Hat was announced. So it can't have been a reason at the time. Though I suspect it's a happy circumstance now for EQT/SUSE.

Can someone explain the business model of suse and redhat ? Is it selling open source software ? How can that be worth billions ? How do you do practically ? Direct sales to enterprise !? Is there a lot of configuration and extra feature implementation, or does most enterprise run "vanilla" ? How do you brak into this market ?

Selling support

Business models of all big software companies are identical and have not changed in decades:

You buy licenses, renew them yearly. With various levels of premium support to select from.

Subscriptions, not licenses.

Subscriptions give you access to different level of support, and access to updates.

Expensive support contracts, and in case of SUSE, some add-on software channels that come with the contracts (SLES for SAP, High Availability, Containers, etc.) being locked away behind paywall.

"Few companies have changed hands as often as Suse and yet remained strong players in their business."

"Strong players"? What quantitative evidence exists demonstrating that SuSE has any more than a marginal and insignificant share of server or desktop installs?

I found this interesting:

"Not that surprising, SLE POS is one of SUSE's best selling products

For example, in the states, 70% of Fortune 100 merchandisers, speciality retailers, and food and drug stores use it

Including a few big names like Walmart"


You picked a strange hill to make your stand on.

Its journalistic verbiage, what do you want them say? Ultimately its been bought for $2.5billion, if it were a startup it would be a unicorn, it seems to be moving forward, what more do you want.

There are probably many companies that have changed hands that many times also.

The market SuSE is in is not "server or desktop installs", it is "support contracts/subscriptions for server or desktop installs", and they have to have a non-negligible share of that market, given that Wikipedia claims they have 1400 employees, which is more than 10% of Red Hat's number.

There is some more data in this article: https://fosspost.org/analytics/insights-on-redhat-suse-the-m...

Interestingly it claims 750 employees for SuSE in 2016; perhaps the business entity sold today includes some departments that were elsewhere in the org-chart in 2016?


While I'm not particularly a fan, it's important to make clear the company SUSE (that is, SUSE Linux GmbH) has two distributions. It funds the community-maintained openSUSE distribution, but also has SUSE Linux Enterprise Server (SLES).

openSUSE has a regular ~2 year support window and is probably mostly used by enthusiasts. The commercial SLES product has a 10 year support window (with paid extension for another 3 years).

It's a paid product, but 10+ years of software updates for a stable Linux platform is something that's pretty rare to find, and certainty a great option for certain products. Hopefully I'm corrected on this, but RedHat tends to be the only other vendor supporting Linux distributions for such a long period of time?

Cannonical does 5 years for lts for everyone , likely more for enterprise customers

Shuttleworth announced in November 18.04 will receive 10 years of support as well.

This applies to Ubuntu Core only: https://www.ubuntu.com/core

Standard Ubuntu still has five years: https://www.ubuntu.com/about/release-cycle

Suse has been pushing btrfs, zypper and other neat tools forward for years. IBM/Toshiba's 4690 POS (used in all the Kroger stores, Fred Meyers, etc), Walmart's in house POS, OfficeMax/OfficeDepot and numerous others use Suse as a base for their point of sale system, deli scales, and other edge applications.

I'm relatively certain 4690 OS is a independent, proprietary OS or at least lacking the Linux kernel, which is pretty important.

IBM stopped development of 4690 OS in 2005, porting all their software to Suse. Toshiba bought the point of sale business from IBM in 2012, but any 4690 based installs are a legacy from the early 2000s, as SuperPOS ACE is what IBM (now Toshiba) has been pushing for the past decade.


Yes, SUSE ("IBM IRES" in this case) is a completely different OS from 4690 and nothing on the wiki suggests otherwise.

From my experience it's been very reliable and yast has made admin changes very easy.

Yup. We use it as well. Quite reliable, easy to upgrade without issues (unlike Ubuntu), easy to configure with yast. I was missing some packages we use for deployment and backup and openSuSE build service came in handy for creating a custom repo. The only issues we had were with old HP hardware. LEAP 15 even has a system that lets you rollback updates but it only works with btrfs (uses fs snapshots) which we don't use out of caution.

Was my first distro too and yes, it was a nightmare. my personal favorite was the gui application to configure X11 and the graphics driver, every failure whould put you back into console sifting through xml configs.

but i've heard good things since.

  it was a nightmare
On some of my HP/Compaq laptops of the time, it was the only major distro that installed right out of the box.

Isn't/Didn't Suse use to be one of the big competitors of RedHat in the OS market for enterprises?

Still a big competitor, more popular in Europe. I managed several high performance clusters that ran on SLES.

It still is!

I don't see a problem with having more actively developed and supported distros to pick from.

It was my first distro when I picked up some installation CDs for it at Microcenter. I never had any issues.

what does SUSE even do in the age of kubernetes?

They work on the core container engine stuff. Aleksa Sarai (who posts in here every once in a while as "cyphar") is the maintainer of runC, which is core to most popular container runtimes (containerd/docker, podman, cri-o/kubernetes, etc.).

They also work on building a platform for trivially deploying and scaling Kubernetes within openSUSE as Kubic[1]. SUSE has a variant of this as the SUSE CaaS platform[2] ("container as a service" platform) product. Kubic is analogous to Fedora CoreOS[3], while CaaSP is similar to RHEL CoreOS that's part of OpenShift[4].

[1]: https://kubic.opensuse.org/

[2]: https://www.suse.com/products/caas-platform/

[3]: https://coreos.fedoraproject.org/

[4]: https://blog.openshift.com/openshift-4-install-experience/

In addition to all of the container work we do, we also still maintain a distribution and do a tonne of upstream work. While people seem to have forgotten about the purpose of distributions with the whole container thing, they still provide a few things:

* Continued upstream development of core system software.

* A host system that is stable and up-to-date (which you need to run your container manager).

* Container images which have up-to-date software, timely security patches, and reasonable security policies.

Distributions are the only members of the ecosystem which actually solve these problems in any meaningful way. It's a shame that we seem to be boring (or even vestigial) at this point despite all of this new software still requiring our work in order to actually run.

Richard Brown gave two talks at FOSDEM 2019 about Kubic, as well:


thank you, never heard of Kubic until now

Sorry but your comment reveals more about the filter bubble you (or maybe we all actually) are in than reality.

While I haven't used Suse for a while in personal projects, Suse is basically the only alternative to Red Hat enterprises have for running Linux with support for 10 years or more from a company deeply involved in kernel development. A pen testing study recently (can't find the link atm) reported that only RHEL and SLES had properly setup ASLR, selinux/AppArmor and other security features.

When Red Hat is fully absorbed into IBM, there might be many customers not willing to make deals with IBM, so I'm guessing the investor sees serious growth for Suse's enterprise support business.

Now if only they shipped a distro without systemd and more POSIXly ...

You do realize that "cloud" doesn't mean the services are running in water vapour up in the sky, right? There are real computers with real operating systems running beneath all those trendy virtualization concepts.

One of those operating systems is SUSE Linux.

I'm amazed that Suse has $2.5B to buy themselves. It's a horrible OS to work with. Anything not based on Apt as a package manager is a nightmare. What does Suse bring to the table other than pain? It's not as easy as Ubuntu, and not as well-known or well-supported as RedHat (which is also a pain in the ass, but at least a well-known pain in the ass). Who even uses Suse?

Suse is rock solid and, for a time, was basically the Redhat of Europe.

I still find places in the UK that run it but it’s less common now than it used to be.

I completely disagree with your point that anything not based in apt is horrible to work with.

(disclaimer: SUSE employee and user)

I get the impression that people using Debian and their derivatives seem to live under a rock regarding other distributions.

I would trust zypper dup hundred times over apt-get dist-upgrade. This gets amplified by the fact that openSUSE Tumbleweed (rolling distro), is much better auto-tested before being published every week, than my experience with sid was (tested by their end-users).

Today's Redhat dnf package manager is based on SUSE's dependency solver (https://github.com/openSUSE/libsolv), and both package managers provide a very nice user experience.

People seem to be still confused about SUSE "using YaST". YaST is just a front-end powered by libzypp (https://github.com/openSUSE/libzypp), like zypper is.

You only think apt is good because you're used to it and probably haven't used it very extensively. Portage is clearly a better package manager, have you used that?

Apt was the first package manager for a widely-used distribution (Debian) that handled dependencies well. I'm certainly biased, and haven't used everything available, but I have to be realistic about Linux distros in a professional environment. Your options at work are RPM-based distros (RedHat, CentOS, SuSE) or Apt-based distros (Ubuntu, and to a lesser extent Debian).

I'd never heard of Portage. Google indicates that it's Gentoo's package manager. It may be fantastic, but Gentoo is a hobby OS. Companies aren't going to use Gentoo. I'm not going to use Gentoo, because any benefits it has are going to be minimal, and it's never going to help my career.

There's a vast array of software out there that might be great for specific purposes, but I'm not going to swim upstream. So, while Gentoo and Portage may be fantastic, they're simply not on my radar, and certainly not on the radar of anyone I may work for.

I'm not disagreeing with you; we're just talking about completely different worlds at this point.

>Google indicates that it's Gentoo's package manager. It may be fantastic, but Gentoo is a hobby OS. Companies aren't going to use Gentoo. I'm not going to use Gentoo, because any benefits it has are going to be minimal, and it's never going to help my career.

>So, while Gentoo and Portage may be fantastic, they're simply not on my radar, and certainly not on the radar of anyone I may work for.

Google's Chrome OS is a Gentoo fork.

>Anything not based on Apt as a package manager is a nightmare.

Portage is not nightmare. Can apt run multiple unrelated installs in parallel?

> Anything not based on Apt as a package manager is a nightmare.

In my experience pacman, APK, OpenBSD's ports and packages, and Slackware's pkgtools are all as good as or better than apt in various ways, with the caveat that Slackware doesn't do dependency resolution, which most Slackware users actually see as an advantage.

To put it another way, there is no perfect package manager, there is only the tool that works best for you. The above tools are what I feel to be the superior package management schemes in no particular order; apt has left me hanging way too many times to be comfortable with it on a production machine. For others (like you) the opposite may be true. Regardless, your statement only applies to you and shouldn't be presented as the gospel truth.

I personally haste to use apt and for me anything is unacceptable other than pacman based distributions.

What's wrong with non-apt?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact