Quick question - do you know if their binary logs are indexed? In BSD is/was quite common to use append-only extended attribute to critical logs when stored locally, to avoid tampering. Is this possible with systemd?
Awww. I'm glad you found time during all your rockstar full stack development to work out that good management is a myth.
Awww. I'm glad you love MBA-types so much. The biggest project I've worked on is an astounding success because it was created without "managers". And while I certainly don't make millions off it, the company probably generates more than a billion in revenue from it each year. And its not like we had a tight budget or didn't have staff available.
I recommend your next step should be to start a company where you don't bother hiring those non-existent good managers - I'm sure you'll be a millionaire in no time.
Yah sure. Lets have a bunch of MBAs discussing acceptance criteria for stuff they have no clue how to work and *if* it fullfills the business requirements. Lets have software rewritten every 6 months because project managers, product owners and most of the upper layers focus on deliver and not on maintainability. Lets do it your way - after all, there is no chance other people have different experience than yours, right?
Nope, worked with them, dismissed several of them because their behaviour was detrimental to the team (one got sacked because he went and told the book keepers that he was more important than they were and should do what he said).
That basically demonstrates you've worked with assholes, not rockstar devs. Its not the same thing.
The ones who actually make the team better dont consider themselves to be rockstars.
So, you never worked with a guy that is somewhat difficult to manage, but has above-average productivity AND is a problem-solver for your team? I'd say you have pretty limited experience. Or bad luck.
This is how they like to imagine they are, but not what they're like in reality. In reality they are childish and petulant. If their authority and awesomeness is not recognised they will make everyone else's life hell until it is.
You seem to have a pretty strong opinion about people you never met - from your own experience. I'd suggest that is the issue: you do have a comfort zone regarding managing devs, and answering defiance (or what you perceive as defiance) is outside of that zone. That is actually your "problem" (or characteristic), not the guys you get onboard.
Largest organisation I worked for in that capacity was 80 staff with 20 developers (most in a consulting capacity)
Freelancers? You're telling me you managed freelancers and, gosh, they didn't delivered as expected?
I had a pair of senior devs who could keep the team together and moving and were great at it, I considered it my job to keep things out of their way so they could do their jobs.
So, apparently you did have rockstars. To a point where you wouldn't even interfere. See?
This is not the job of a single, lone, "superstar" programmer but of a fluent, experienced, team of professionals who know how to work together.
Having been the "lone superstar" on a couple of global projects, I can assure you there is a time and a place for everyone. While I don't agree with the stupid "10x productivity", a "rockstar engineer" will save you money and trouble, and make sure the project meets or exceeds the criteria. If the only guys you've worked with are unsufferable code monkeys (and usually existing codebases tend to get *smaller* with me, not bigger), its your problem. And, in my case, having both business experience on the sector and being a full-stack engineer, I spotted and fixed problems in the specifications that would take weeks to identify and fix, saving time, trouble and money. I'm not the most pleasing guy to work with, and as most guys like me, I'm not really attracted to long-term maintenance/bugfixing jobs, but I'd say there is definitely a place for people like me in most projects.
The key to fast project delivery is good management and perceptive staff selection
Good management is a myth, unless your management stack is comprised of individuals smarter than you on the specific field. It can help a lot (and I've worked with wonderful management), but that's it. Its not a silver-bullet.
Looking for a superstar programmer as some sort of silver-bullet is both naive and doomed to failure
Having a superstar programmer in your team will help push the envelope and inspire younger programmers. There is always so much to learn, and not everything is about lines of code. Its about tools, processes, thinking outside the box. If you measure the rate of a programmer by its code deliveries, well... you have flawed metrics and in the end you get what you paid for.
If RH never made another release, there would be similar disruption.
I doubt it. There is an OSS-based fork with many of the features RH clients actually use. Its called CentOS, and that alone is what drives the RH ecosystem. Very few customers actually *run* RH, and those who do usually do it for compliance and certfication issues (think about SAP and Oracle installs).
That doom theorycrafting is irrelevant to my question.
You made no question.
That's also irrelevant. They are distros from a business standpoint. CentOS being interchangeable with Fedora since forever. How they came to be is a footnote.
CentOS has a user base that is probably an order of magnitude bigger than RH. So proportionally, you're just agreeing with me. They are small fish.
My question was about relative usage and some way to measure that metric other than guesswork, as a challenge to the assertion that RH is "a marginal player". systemd adoption in RH is mentioned in 100% of the "discussions" on the topic. So someone here is showing bias.
RH is a company. There are already established metrics for that. Its a mix between revenue and net profit, usually. And as I've shown you, RH is a marginal player.
If you were going to address the issue in an objective manner, you might note Debian, tends to identify itself when you run fingerprinting on servers (e.g. Apache and Nginx). Debian tends to be the most common identifier! Nobody believes the bulk of the responses (with no OS identifiers) are all non-RH (some will be slack, some debian, some gentoo, whatever), so that's an interesting metric that isn't definitive.
No one cares about webservers, and I've never seen anybody buying linux (so, RH customers) to run Nginx. Webservers are a commodity, and can be run in whatever you want. There is nothing that makes debian or linux by itself more interesting than the rest. Many new-breed sysadmins are used to ubuntu, thats why they choose debian.
think I understood completely. Attempting to derail into some form of "RH can be replaced" discussion, is of no interest to me.
RH *can* be easily replaced. Its one of those companies that have no real added value in their stack, other than support. See the evolution of sales of RH and compare it to other players, and you'll easily find out that.
Calling RH a marginal player is simply disingenuous, as of today.
The numbers are there. Like it or don't - it is a marginal player.
Also, SAP and CA's sales income is irrelevant for comparison here, since they aren't in the operating system business.
No, they rely on operating systems as a critical part of their business. The day that operating system is a liability, well.. then business sizes matter a lot.
If RH made a change that impacted Oracle or CA - Oracle or CA would have to adapt.
Are you kidding me? RH is the cheap stuff. Oracle pitches the Solaris solutions for the high-end customers (and that is the big reason they bought Sun), and they maintain their own RH-based fork - Unbreakable Linux. So basically, a prime example of what I said.
Its not a metric. It just means that many of the mentioned companies can buy out RH with their lunch money, if some disruptive change challenges their business model. And as an example, Oracle DB is *not* certified to run on Debian (it basically means you're even more dead in the water in case of problems). If RH introduced changes that disturbed the Oracle ecosystem, size does matter. And RH is small fish.
Please re-read my comment. I don't think you understood it.
The day RH choices disturb any big company from their own ecosystem, they will be eaten alive. And I've shown you the numbers that prove it.
RH *is* a business, Debian is a community effort. Business get bought all the time. Community efforts live while the community is healthy. Debian is quite common because is a simple and easy to install as a barebones version. Not because its better somehow. Its the same dogfood everybody else eats. Btw, and since you mention EC2, imagine Debian adopted something that would disrupt EC2 funcionalities - it would be replaced quite quickly.
I continue to hear this and see absolutely no evidence of it. I see evidence to the contrary, in the US, India and Europe, over the last 20 years.
Generally, it's RPM/RH that is first listed. It's not alphabetical. This isn't because they are lucky. The simple explanation is that RH is the most frequently used and therefore put at the top as a simple matter of UI layout (most common choices go to the top of a list, within reason).
This is a fun game, pick me a list that shows more Debian love!
If you rank popularity from the links of the packages, you provably are missing something. I'm a FreeBSD/OpenBSD user, some of the development of Nginx was done in FreeBSD, and you don't even see packages for it in your list. See the flaw in your methodology?
It still is a huge limitation, as you cannot easily sync a local dataset with a remote one.
I assume that's why you ignored Red Shift.
RedShift has a limitation of 16TB per node. Its nice, but not really "big data". Its more like RDS on steroids, and I think RDS is "so-so". Also, you either use sync from/to amazon interfaces, or you're stuck with JBDC, so basically the same limitation you mentioned apply.
I recommend against MSSQL not because it's not a good DB
I'm assuming this is based on your extensive MSSQL experience, right?
but because it's cumbersome to work with outside of the Microsoft ecosystem.
Well, at least MS has half-decent tools for it. Other than Oracle, they're the only player with a decent GUI interface.
You mainly interface with it using ODBC and that's a pain outside of Windows.
Or JBDC. It depends. Its not really Microsoft's fault if it doesn't work in your environment, is it?
You're stuck with windows boxes on the back end AND on the front end.
You just summarized 2/3rds of the corporate world.
You could start your project with Postgres and find out why you're unhappy with it and plan for a migration to something which is better for you post-hoc: Don't write SQL procs, and don't weave your SQL through a whole lot of code.
You could, and then figure out how to integrate it with your environment. And *WHICH* ODBC driver to choose. So, the pain you just described previously, its right there. With a half-assed, subpar connection driver.
The only OSS solution that comes somewhere *near* what MSSQL does is PostgreSQL, and its a second-class citizen in Windows. And even PgSQL is easily suprassed when looking at features and replication options.
Note: I'm a huge PostgreSQL fan, to the point of writing C# applications with the native PostgreSQL driver without LINQ support.I'd take PostgreSQL over MSSQL everyday of the week if reporting tools, support, features, replication and integration doesn't matter. But saying that MSSQL is bad, is just a silly mantra.
(...) in a FOSS environment the USERS get a say in things?
Well, yes and no. Developers also use their own dogfood, and they're a non-trivial percentage of the userbase.
We know what is good for you, you filthy peasant" top down management bullshit that treats the users as ignorant children
As you frequently post lets-call-it-pro-microsoft-stuff (sometimes you hit the head of the nail, sometimes a finger), it's not too much different from what Microsoft (and other companies) try to do. The idea that "users will get used to" was big with Windows XP, and it was a major success. They try to repeat it ever since and fail miseably (although, I look at the metro interface I personally don't use (as an old fart windows user, I demand some consistency) and I think that with a couple of changes, its friggin awesome - light years ahead of what those guys at Canonical are doing.
Compare this to the Linux side where Red Hat is basically pulling a coup and taking over the whole show
Debian has no vote on - nor does it care - about Red Hat policies.
If RH truly thinks their way is better let it sink or swim on the merits, not because all dissent was crushed.
RH is a minority in the landscape where systemd affects systems. When RH (which is, both in business model and revenue, a small player in the IT panorama) starts to hinder, lets say - Oracle or CA interests, they'll be bough or put in line. The widespread adoption of systemd will force other, smaller distros to adopt it, because of software dependency - we've already seen this this Gnome3 and BSD support. It is perfectly normal that a sane person decides to leave because the project he dedicates his time to is no longer the project he joined, but a pawn of corporate interests (Debian is the "de facto" independant linux distro source, and most of the popular distros extend from it).
In short, this is not about the users. It never was. Its about power plays and politics. In Microsoft world, they'd be guidelines and early adoption sneak previews.
...And yet you probably don't have polyo because some guy decided to give the cure to the world.
Have a look at your software stack, and the hardware that follows, and all the people that innovated without thinking about money first. Whatever you are today, you owe to those people. And most of them don't really care if you made yourself rich or not - if you cannot conceive this concept, I pitty you. No money in the world will fill that hole.
Not really, no. Theoretically, they are possible (assuming same source IP and predictable sequence numbers after handshake). In practice, it doesn't really work that way - because in the end, you may be sending the ack packet from the wrong interface, and mess the state table there:)
One caution is that Windows is not as secure an OS perhaps because
there is a rich set of stuff that is darn hard to replace or eliminate.
I haven't seen one single landline direct-connection to the internet since the dialup/adsl days. Most consumers will have a router. The only exception is 3G/4G adapters, but the topic is about firewalling. And unless you're running a DPI appliance to check for binary malware, you're getting those in your windows machines anyway.
A FreeBSD or Linux based firewall+VPN system can be pruned to an astoundingly
short list of services and binaries
As can Windows. And you can also take the easy approach of just closing any external port besides the VPN, leaving only potential attacks on the TCP stack and the VPN layer. I actually find funny people that use firewalls on unix systems "as a checklist item"; Most systems don't even require firewall if properly configured. But yeah, lets badmouth windows and forget the ton of distros that allow remote root login via ssh *by default*.
You open up a good context to make the point that a user should use what
they know best. If the poster knows how to manage one system and not
the other then the best answer for that user is obvious.
No. If the user knew what was best - or at least the options available - he woudn't be asking this. Having guys following tutorials on the internet to configure stuff is not my idea of "secure", and he'd probably be better buying a dedicated appliance with a nice gui interface.While realizing that you exposed something from the internal system or used a weak password for root after your whole network was compromised does have its educational value, it is a dreadful experience for a non-unix nerd.
Full blown Win-Server
software that can get the job done costs more than the hardware.
No, not really. Windows has the easiest internet-sharing and vpn configuration wizard you'lll find. And its not half bad, but...
The above is a rather nice little box. At half this price I would buy two.
I have an equivalent box, Instead of pfSense (which, besides the gui and the easy VLAN setup, is a crappy system for everything else), I run FreeBSD 9.2. And I use it everyday to tunnel into my windows machines with RDP via SSH:)
ZFS is a big monolithic package that does everything, much like Microsoft Word or Outlook. ZFS is more in the Microsoft tradition.
Well, that is well within the Unix tradition. ZFS is a *kernel* module, not a userland application. Just because the cli interface is comprised of 2 commands, it doesn't mean its monolithic. Its as monolithic as ifconfig and other complex utilities.
And I'd take anyday the zfs/zpool command format over the lvm ugly mess.
If no major motherboard manufacturer even cares about niche market then I would ask you to explain all of the boards that are targeted towards multi-GPU setups and overclocking which are both small minority niche markets.
You say they are small. They aren't. They are the beef of the desktop market - hardcore gamers.
I beg to differ on this one as well. Just look at what Linux admins or really anyone with Linux skills gets paid compared to Windows admins. Not a single person that I know who really knows tech buys Apple, or Microsoft products (other than Windows for gaming)
You can differ as much as you want. Competent Linux admins and Windows admins are paid about the same. Less on the linux side for most run-to-the-mill LAMP setups, more on Windows side for enterprise. You pay for competency first, and any above-average sysadmin will be proficient in several systems, not only Linux. But in the end, techies are a minority, way smaller than gamers. And just because you're a sysadmin, it doesn't mean you can distinguish good gear from bad gear. Lots of techies I know do use Linux. And FreeBSD. And DragonFlyBSD. And Windows. A good techie isn't usually a one trick monkey. The sooner you learn it, the faster you'll grow.
People who know a little more might buy things like ECS or Foxcon motherboards that in my experience are not worth even thinking about as they tend to be less compatible and have lower quality components like capacitors that are being pushed to their limits which causes them to break down much more quickly.
Actually, Foxconn are usually replicas of Intel desktop silks. When a chipset is released, it is often accompanied with reference schematics that are - basically - a skeleton version of the reference board for that chipset. Foxconn usually mimics it to a point where you can hold both motherboards and they seem to differ only in color. So they usually are as compatible as one can be.
Regarding "pushing capacitors to the limit", isn't really about that. Its about electrolyte degradation. And this can happen with any major manufacturer, as they don't control every step of the supply chain.
People that know what they're buying are buying ready-made workstations from Dell or HP or Apple, or building it with Tyan, SuperMicro or similar gear. Coincidentally, both Dell and HP are huge players in the server market, and they use Foxconn factories.
I actually have used desktop components for servers quite a lot. I do make sure there is redundancy and for a small business they really do not need anything more. Also in my experience good quality desktop components are just as stable and last just as long as server components. Besides who cares if the system lasts for 5 years or 10 years when it should be considered too slow to be useful after 3 years?
Yah, that shows. You're "that" kind of guy. Let me ask you, assuming you're running eg. databases on those servers, what happens when a bit is flipped on in memory and a write operation commits 0x10FE credit instead of 0xFE? Your redundant system will replicate this and silently propagate the error. Or when a block is misread from a single disk instead of using parity check? Are your clients aware that this can happen? Have you explained it to them?
I have run many different Windows and Linux servers and have worked for hosting providers that host 1000's of websites and other applications on both and I can say from experience that Windows uses a lot more resources, is much slower and is much less stable than Linux, and in many cases Linux is quicker and easier to get setup and running, although not always quite as straightforward as Windows.
So, you have Windows and Linux experience on a very narrow field. Good for you. I can actually setup an OpenBSD server way faster than you can install most Linux distros, does it mean its a good replacement for every workload? Not
Smaller companies often have old desktops running as their "servers", no raid (or using the crappy bios fakeraid), no backups, no redundancy etc.
Smaller companies often have no servers and have everything online, or have a in-house NAS and a bunch of desktops. This isn't the nineties anymore. Some corner shops may still have a couple of desktops doubling as servers (yah, I've seen it), but it is not that common.
Lots of cheaper servers are also based on desktop boards, and lots of budget hosting companies use such systems.
Just because they are in a rackmount case, it doesn't make them "servers". And most providers describe in detail the hardware, and will give you explicit option for an entry-level server solution - you get what you choose to pay for. If you're dumb enough to get an i7 "server" with 32GB of RAM for database work, its your problem, not theirs.
Most desktop gear isn't even designed for a 24/7 operation, let alone having to support the cpu running at full capacity and indefinite amount of time. Desktop gear is not designed, both from a thermal and electrical perspective, for this kind of operation.
I like to use ECC even on the desktop, and yes there are ways to do it. At a cost.
Just because you use entry-level server gear/workstation gear for desktop work, it doesn't make it desktop:) And the cost isn't that big of a difference - a small entry-level xeon workstation/server isn't that much more expensive than the desktop gear. The problem is, most "regular" consumers aren't even aware of this, because the corner shop and the shopping mall sells computers that look like a christmas tree, not workstations.
And you think with the low margins the manufacturers have these days, they can do without that share?
Unfortunely, yes. No major motherboard manufacturer even cares about niche market. And the IC manufacturers, they don't really care, either.
Also people using desktop Linux are typically in the higher income levels and can not only pay for quality
Higher income buyers are buying trendy Apple, Andoid tablets and Microsoft laptops, not linux workstations.
they can recognize it, unlike the sheep
No, they just don't care about that. But you do get the smugness of the illusion that the manufacturer uses fairy dust instead of building it like everyone else.
Wolves are always a minority.
Now, you're just assuming stuff. I'd say wolves are quite the majority of animals in wolfpacks, and the major ingredient in wolf stoo.
What you are also completely forgetting is that a lot of these will actually run as servers. You know, because Linux does well as server
Who is using COTS desktop boards on servers? Traditionally, Intel desktop cpu lines do not support ECC memory. And you talk like there is no option for servers besides Linux.
You know, because Linux does well as server, quite unlike Windows
I assume you speak from experience. I'd blame it on the sysadmin, not the operating system.
But you would not know or understand that.
Get out of the basement sometimes. Try to vent out at least some of that frustration of yours.
In the early days all of Germany's neighbors were weak in comparison, they'd all been through WW1 also.
I don't really see where you got that idea. The british empire had what was probably the most well-oiled war machine at the time - and Germany gave them a run for their money.
No one really wanted to get into a war and were slow to react.
If you mean "leaders were well aware of the consequences of war in your own backyard because it had happened a decade and a half before", yeah. No redneck reasoning here. And when the Americans realized this, they bombed Hiroshima and Nagasaki. They've witnessed the horrors of the invasion of China and what is now Korea a couple of years before. (Hint: google "nanking massacre" - it makes the german seem inefficient by comparison).
If Germany had just grabbed their "lebensraum" and stopped they probably could have kept it.
You seem to forget that WWII was also fought in Africa. And Asia. Germany was fighting every major country on Earth at the time (with the exception of Japan), and had taken control of France and Netherlands, both with colonies in Africa.
Slashdot Mirror
Quick question - do you know if their binary logs are indexed? In BSD is/was quite common to use append-only extended attribute to critical logs when stored locally, to avoid tampering. Is this possible with systemd?
Awww. I'm glad you found time during all your rockstar full stack development to work out that good management is a myth.
Awww. I'm glad you love MBA-types so much. The biggest project I've worked on is an astounding success because it was created without "managers". And while I certainly don't make millions off it, the company probably generates more than a billion in revenue from it each year. And its not like we had a tight budget or didn't have staff available.
I recommend your next step should be to start a company where you don't bother hiring those non-existent good managers - I'm sure you'll be a millionaire in no time.
Yah sure. Lets have a bunch of MBAs discussing acceptance criteria for stuff they have no clue how to work and *if* it fullfills the business requirements. Lets have software rewritten every 6 months because project managers, product owners and most of the upper layers focus on deliver and not on maintainability. Lets do it your way - after all, there is no chance other people have different experience than yours, right?
Nope, worked with them, dismissed several of them because their behaviour was detrimental to the team (one got sacked because he went and told the book keepers that he was more important than they were and should do what he said).
That basically demonstrates you've worked with assholes, not rockstar devs. Its not the same thing.
The ones who actually make the team better dont consider themselves to be rockstars.
So, you never worked with a guy that is somewhat difficult to manage, but has above-average productivity AND is a problem-solver for your team? I'd say you have pretty limited experience. Or bad luck.
This is how they like to imagine they are, but not what they're like in reality. In reality they are childish and petulant. If their authority and awesomeness is not recognised they will make everyone else's life hell until it is.
You seem to have a pretty strong opinion about people you never met - from your own experience. I'd suggest that is the issue: you do have a comfort zone regarding managing devs, and answering defiance (or what you perceive as defiance) is outside of that zone. That is actually your "problem" (or characteristic), not the guys you get onboard.
Largest organisation I worked for in that capacity was 80 staff with 20 developers (most in a consulting capacity)
Freelancers? You're telling me you managed freelancers and, gosh, they didn't delivered as expected?
I had a pair of senior devs who could keep the team together and moving and were great at it, I considered it my job to keep things out of their way so they could do their jobs.
So, apparently you did have rockstars. To a point where you wouldn't even interfere. See?
This is not the job of a single, lone, "superstar" programmer but of a fluent, experienced, team of professionals who know how to work together.
Having been the "lone superstar" on a couple of global projects, I can assure you there is a time and a place for everyone. While I don't agree with the stupid "10x productivity", a "rockstar engineer" will save you money and trouble, and make sure the project meets or exceeds the criteria. If the only guys you've worked with are unsufferable code monkeys (and usually existing codebases tend to get *smaller* with me, not bigger), its your problem.
And, in my case, having both business experience on the sector and being a full-stack engineer, I spotted and fixed problems in the specifications that would take weeks to identify and fix, saving time, trouble and money. I'm not the most pleasing guy to work with, and as most guys like me, I'm not really attracted to long-term maintenance/bugfixing jobs, but I'd say there is definitely a place for people like me in most projects.
The key to fast project delivery is good management and perceptive staff selection
Good management is a myth, unless your management stack is comprised of individuals smarter than you on the specific field. It can help a lot (and I've worked with wonderful management), but that's it. Its not a silver-bullet.
Looking for a superstar programmer as some sort of silver-bullet is both naive and doomed to failure
Having a superstar programmer in your team will help push the envelope and inspire younger programmers. There is always so much to learn, and not everything is about lines of code. Its about tools, processes, thinking outside the box. If you measure the rate of a programmer by its code deliveries, well... you have flawed metrics and in the end you get what you paid for.
What 14 years? http://web.archive.org/web/199...
Yah, you are actually right. I didn't see it. My mistake.
If RH never made another release, there would be similar disruption.
I doubt it. There is an OSS-based fork with many of the features RH clients actually use. Its called CentOS, and that alone is what drives the RH ecosystem. Very few customers actually *run* RH, and those who do usually do it for compliance and certfication issues (think about SAP and Oracle installs).
That doom theorycrafting is irrelevant to my question.
You made no question.
That's also irrelevant. They are distros from a business standpoint. CentOS being interchangeable with Fedora since forever. How they came to be is a footnote.
CentOS has a user base that is probably an order of magnitude bigger than RH. So proportionally, you're just agreeing with me. They are small fish.
My question was about relative usage and some way to measure that metric other than guesswork, as a challenge to the assertion that RH is "a marginal player". systemd adoption in RH is mentioned in 100% of the "discussions" on the topic. So someone here is showing bias.
RH is a company. There are already established metrics for that. Its a mix between revenue and net profit, usually. And as I've shown you, RH is a marginal player.
If you were going to address the issue in an objective manner, you might note Debian, tends to identify itself when you run fingerprinting on servers (e.g. Apache and Nginx). Debian tends to be the most common identifier! Nobody believes the bulk of the responses (with no OS identifiers) are all non-RH (some will be slack, some debian, some gentoo, whatever), so that's an interesting metric that isn't definitive.
No one cares about webservers, and I've never seen anybody buying linux (so, RH customers) to run Nginx. Webservers are a commodity, and can be run in whatever you want. There is nothing that makes debian or linux by itself more interesting than the rest. Many new-breed sysadmins are used to ubuntu, thats why they choose debian.
think I understood completely. Attempting to derail into some form of "RH can be replaced" discussion, is of no interest to me.
RH *can* be easily replaced. Its one of those companies that have no real added value in their stack, other than support. See the evolution of sales of RH and compare it to other players, and you'll easily find out that.
Calling RH a marginal player is simply disingenuous, as of today.
The numbers are there. Like it or don't - it is a marginal player.
Also, SAP and CA's sales income is irrelevant for comparison here, since they aren't in the operating system business.
No, they rely on operating systems as a critical part of their business. The day that operating system is a liability, well.. then business sizes matter a lot.
If RH made a change that impacted Oracle or CA - Oracle or CA would have to adapt.
Are you kidding me? RH is the cheap stuff. Oracle pitches the Solaris solutions for the high-end customers (and that is the big reason they bought Sun), and they maintain their own RH-based fork - Unbreakable Linux. So basically, a prime example of what I said.
Its not a metric. It just means that many of the mentioned companies can buy out RH with their lunch money, if some disruptive change challenges their business model. And as an example, Oracle DB is *not* certified to run on Debian (it basically means you're even more dead in the water in case of problems). If RH introduced changes that disturbed the Oracle ecosystem, size does matter. And RH is small fish.
Please re-read my comment. I don't think you understood it.
The day RH choices disturb any big company from their own ecosystem, they will be eaten alive. And I've shown you the numbers that prove it.
RH *is* a business, Debian is a community effort. Business get bought all the time. Community efforts live while the community is healthy. Debian is quite common because is a simple and easy to install as a barebones version. Not because its better somehow. Its the same dogfood everybody else eats. Btw, and since you mention EC2, imagine Debian adopted something that would disrupt EC2 funcionalities - it would be replaced quite quickly.
I continue to hear this and see absolutely no evidence of it. I see evidence to the contrary, in the US, India and Europe, over the last 20 years. Generally, it's RPM/RH that is first listed. It's not alphabetical. This isn't because they are lucky. The simple explanation is that RH is the most frequently used and therefore put at the top as a simple matter of UI layout (most common choices go to the top of a list, within reason).
Here: http://investors.redhat.com/re...
You can also see the report from CA, where they make more in a quarter than a full year of RH: http://www.ca.com/us/news/pres...
This is a fun game, pick me a list that shows more Debian love!
If you rank popularity from the links of the packages, you provably are missing something. I'm a FreeBSD/OpenBSD user, some of the development of Nginx was done in FreeBSD, and you don't even see packages for it in your list. See the flaw in your methodology?
It's what you'd find that you didn't like.
It still is a huge limitation, as you cannot easily sync a local dataset with a remote one.
I assume that's why you ignored Red Shift.
RedShift has a limitation of 16TB per node. Its nice, but not really "big data". Its more like RDS on steroids, and I think RDS is "so-so". Also, you either use sync from/to amazon interfaces, or you're stuck with JBDC, so basically the same limitation you mentioned apply.
I recommend against MSSQL not because it's not a good DB
I'm assuming this is based on your extensive MSSQL experience, right?
but because it's cumbersome to work with outside of the Microsoft ecosystem.
Well, at least MS has half-decent tools for it. Other than Oracle, they're the only player with a decent GUI interface.
You mainly interface with it using ODBC and that's a pain outside of Windows.
Or JBDC. It depends. Its not really Microsoft's fault if it doesn't work in your environment, is it?
You're stuck with windows boxes on the back end AND on the front end.
You just summarized 2/3rds of the corporate world.
You could start your project with Postgres and find out why you're unhappy with it and plan for a migration to something which is better for you post-hoc: Don't write SQL procs, and don't weave your SQL through a whole lot of code.
You could, and then figure out how to integrate it with your environment. And *WHICH* ODBC driver to choose. So, the pain you just described previously, its right there. With a half-assed, subpar connection driver.
The only OSS solution that comes somewhere *near* what MSSQL does is PostgreSQL, and its a second-class citizen in Windows. And even PgSQL is easily suprassed when looking at features and replication options.
Note: I'm a huge PostgreSQL fan, to the point of writing C# applications with the native PostgreSQL driver without LINQ support.I'd take PostgreSQL over MSSQL everyday of the week if reporting tools, support, features, replication and integration doesn't matter. But saying that MSSQL is bad, is just a silly mantra.
(...) in a FOSS environment the USERS get a say in things?
Well, yes and no. Developers also use their own dogfood, and they're a non-trivial percentage of the userbase.
We know what is good for you, you filthy peasant" top down management bullshit that treats the users as ignorant children
As you frequently post lets-call-it-pro-microsoft-stuff (sometimes you hit the head of the nail, sometimes a finger), it's not too much different from what Microsoft (and other companies) try to do. The idea that "users will get used to" was big with Windows XP, and it was a major success. They try to repeat it ever since and fail miseably (although, I look at the metro interface I personally don't use (as an old fart windows user, I demand some consistency) and I think that with a couple of changes, its friggin awesome - light years ahead of what those guys at Canonical are doing.
Compare this to the Linux side where Red Hat is basically pulling a coup and taking over the whole show
Debian has no vote on - nor does it care - about Red Hat policies.
If RH truly thinks their way is better let it sink or swim on the merits, not because all dissent was crushed.
RH is a minority in the landscape where systemd affects systems. When RH (which is, both in business model and revenue, a small player in the IT panorama) starts to hinder, lets say - Oracle or CA interests, they'll be bough or put in line. The widespread adoption of systemd will force other, smaller distros to adopt it, because of software dependency - we've already seen this this Gnome3 and BSD support. It is perfectly normal that a sane person decides to leave because the project he dedicates his time to is no longer the project he joined, but a pawn of corporate interests (Debian is the "de facto" independant linux distro source, and most of the popular distros extend from it).
In short, this is not about the users. It never was. Its about power plays and politics. In Microsoft world, they'd be guidelines and early adoption sneak previews.
...And yet you probably don't have polyo because some guy decided to give the cure to the world. Have a look at your software stack, and the hardware that follows, and all the people that innovated without thinking about money first. Whatever you are today, you owe to those people. And most of them don't really care if you made yourself rich or not - if you cannot conceive this concept, I pitty you. No money in the world will fill that hole.
You don't really need it. OpenVPN itself is UDP-based, and everyrhing else is handled by native routing.
Not really, no. Theoretically, they are possible (assuming same source IP and predictable sequence numbers after handshake). In practice, it doesn't really work that way - because in the end, you may be sending the ack packet from the wrong interface, and mess the state table there :)
One caution is that Windows is not as secure an OS perhaps because there is a rich set of stuff that is darn hard to replace or eliminate.
I haven't seen one single landline direct-connection to the internet since the dialup/adsl days. Most consumers will have a router. The only exception is 3G/4G adapters, but the topic is about firewalling. And unless you're running a DPI appliance to check for binary malware, you're getting those in your windows machines anyway.
A FreeBSD or Linux based firewall+VPN system can be pruned to an astoundingly short list of services and binaries
As can Windows. And you can also take the easy approach of just closing any external port besides the VPN, leaving only potential attacks on the TCP stack and the VPN layer. I actually find funny people that use firewalls on unix systems "as a checklist item"; Most systems don't even require firewall if properly configured. But yeah, lets badmouth windows and forget the ton of distros that allow remote root login via ssh *by default*.
You open up a good context to make the point that a user should use what they know best. If the poster knows how to manage one system and not the other then the best answer for that user is obvious.
No. If the user knew what was best - or at least the options available - he woudn't be asking this. Having guys following tutorials on the internet to configure stuff is not my idea of "secure", and he'd probably be better buying a dedicated appliance with a nice gui interface.While realizing that you exposed something from the internal system or used a weak password for root after your whole network was compromised does have its educational value, it is a dreadful experience for a non-unix nerd.
Full blown Win-Server software that can get the job done costs more than the hardware.
No, not really. Windows has the easiest internet-sharing and vpn configuration wizard you'lll find. And its not half bad, but...
The above is a rather nice little box. At half this price I would buy two.
I have an equivalent box, Instead of pfSense (which, besides the gui and the easy VLAN setup, is a crappy system for everything else), I run FreeBSD 9.2. And I use it everyday to tunnel into my windows machines with RDP via SSH :)
ZFS is a big monolithic package that does everything, much like Microsoft Word or Outlook. ZFS is more in the Microsoft tradition.
Well, that is well within the Unix tradition. ZFS is a *kernel* module, not a userland application. Just because the cli interface is comprised of 2 commands, it doesn't mean its monolithic. Its as monolithic as ifconfig and other complex utilities.
And I'd take anyday the zfs/zpool command format over the lvm ugly mess.
If no major motherboard manufacturer even cares about niche market then I would ask you to explain all of the boards that are targeted towards multi-GPU setups and overclocking which are both small minority niche markets.
You say they are small. They aren't. They are the beef of the desktop market - hardcore gamers.
I beg to differ on this one as well. Just look at what Linux admins or really anyone with Linux skills gets paid compared to Windows admins. Not a single person that I know who really knows tech buys Apple, or Microsoft products (other than Windows for gaming)
You can differ as much as you want. Competent Linux admins and Windows admins are paid about the same. Less on the linux side for most run-to-the-mill LAMP setups, more on Windows side for enterprise. You pay for competency first, and any above-average sysadmin will be proficient in several systems, not only Linux. But in the end, techies are a minority, way smaller than gamers. And just because you're a sysadmin, it doesn't mean you can distinguish good gear from bad gear. Lots of techies I know do use Linux. And FreeBSD. And DragonFlyBSD. And Windows. A good techie isn't usually a one trick monkey. The sooner you learn it, the faster you'll grow.
People who know a little more might buy things like ECS or Foxcon motherboards that in my experience are not worth even thinking about as they tend to be less compatible and have lower quality components like capacitors that are being pushed to their limits which causes them to break down much more quickly.
Actually, Foxconn are usually replicas of Intel desktop silks. When a chipset is released, it is often accompanied with reference schematics that are - basically - a skeleton version of the reference board for that chipset. Foxconn usually mimics it to a point where you can hold both motherboards and they seem to differ only in color. So they usually are as compatible as one can be.
Regarding "pushing capacitors to the limit", isn't really about that. Its about electrolyte degradation. And this can happen with any major manufacturer, as they don't control every step of the supply chain.
People that know what they're buying are buying ready-made workstations from Dell or HP or Apple, or building it with Tyan, SuperMicro or similar gear. Coincidentally, both Dell and HP are huge players in the server market, and they use Foxconn factories.
I actually have used desktop components for servers quite a lot. I do make sure there is redundancy and for a small business they really do not need anything more. Also in my experience good quality desktop components are just as stable and last just as long as server components. Besides who cares if the system lasts for 5 years or 10 years when it should be considered too slow to be useful after 3 years?
Yah, that shows. You're "that" kind of guy. Let me ask you, assuming you're running eg. databases on those servers, what happens when a bit is flipped on in memory and a write operation commits 0x10FE credit instead of 0xFE? Your redundant system will replicate this and silently propagate the error. Or when a block is misread from a single disk instead of using parity check? Are your clients aware that this can happen? Have you explained it to them?
I have run many different Windows and Linux servers and have worked for hosting providers that host 1000's of websites and other applications on both and I can say from experience that Windows uses a lot more resources, is much slower and is much less stable than Linux, and in many cases Linux is quicker and easier to get setup and running, although not always quite as straightforward as Windows.
So, you have Windows and Linux experience on a very narrow field. Good for you. I can actually setup an OpenBSD server way faster than you can install most Linux distros, does it mean its a good replacement for every workload? Not
Smaller companies often have old desktops running as their "servers", no raid (or using the crappy bios fakeraid), no backups, no redundancy etc.
Smaller companies often have no servers and have everything online, or have a in-house NAS and a bunch of desktops. This isn't the nineties anymore. Some corner shops may still have a couple of desktops doubling as servers (yah, I've seen it), but it is not that common.
Lots of cheaper servers are also based on desktop boards, and lots of budget hosting companies use such systems.
Just because they are in a rackmount case, it doesn't make them "servers". And most providers describe in detail the hardware, and will give you explicit option for an entry-level server solution - you get what you choose to pay for. If you're dumb enough to get an i7 "server" with 32GB of RAM for database work, its your problem, not theirs.
Most desktop gear isn't even designed for a 24/7 operation, let alone having to support the cpu running at full capacity and indefinite amount of time. Desktop gear is not designed, both from a thermal and electrical perspective, for this kind of operation.
I like to use ECC even on the desktop, and yes there are ways to do it. At a cost.
Just because you use entry-level server gear/workstation gear for desktop work, it doesn't make it desktop :) And the cost isn't that big of a difference - a small entry-level xeon workstation/server isn't that much more expensive than the desktop gear. The problem is, most "regular" consumers aren't even aware of this, because the corner shop and the shopping mall sells computers that look like a christmas tree, not workstations.
And you think with the low margins the manufacturers have these days, they can do without that share?
Unfortunely, yes. No major motherboard manufacturer even cares about niche market. And the IC manufacturers, they don't really care, either.
Also people using desktop Linux are typically in the higher income levels and can not only pay for quality
Higher income buyers are buying trendy Apple, Andoid tablets and Microsoft laptops, not linux workstations.
they can recognize it, unlike the sheep
No, they just don't care about that. But you do get the smugness of the illusion that the manufacturer uses fairy dust instead of building it like everyone else.
Wolves are always a minority.
Now, you're just assuming stuff. I'd say wolves are quite the majority of animals in wolfpacks, and the major ingredient in wolf stoo.
What you are also completely forgetting is that a lot of these will actually run as servers. You know, because Linux does well as server
Who is using COTS desktop boards on servers? Traditionally, Intel desktop cpu lines do not support ECC memory. And you talk like there is no option for servers besides Linux.
You know, because Linux does well as server, quite unlike Windows
I assume you speak from experience. I'd blame it on the sysadmin, not the operating system.
But you would not know or understand that.
Get out of the basement sometimes. Try to vent out at least some of that frustration of yours.
In the early days all of Germany's neighbors were weak in comparison, they'd all been through WW1 also.
I don't really see where you got that idea. The british empire had what was probably the most well-oiled war machine at the time - and Germany gave them a run for their money.
No one really wanted to get into a war and were slow to react.
If you mean "leaders were well aware of the consequences of war in your own backyard because it had happened a decade and a half before", yeah. No redneck reasoning here. And when the Americans realized this, they bombed Hiroshima and Nagasaki. They've witnessed the horrors of the invasion of China and what is now Korea a couple of years before. (Hint: google "nanking massacre" - it makes the german seem inefficient by comparison).
If Germany had just grabbed their "lebensraum" and stopped they probably could have kept it.
You seem to forget that WWII was also fought in Africa. And Asia. Germany was fighting every major country on Earth at the time (with the exception of Japan), and had taken control of France and Netherlands, both with colonies in Africa.