On Sun, 08 Nov 2009 11:11:08 +0000, TomB wrote: > On 2009-11-08, the following emerged from the brain of Snit: >> Hadron stated in post hd5snk$bm...@news.eternal-september.org on >> 11/8/09 12:41 AM:
>>> "DFS" <nospam@dfs_.com> writes:
>>>> TomB wrote:
>>>>> One mistake you make is considering the desktop and server editions >>>>> of Ubuntu as two different distros. They are not.
>>>> Different editions, different distros, different default packages, >>>> different download sites, etc.
>>> And then add the "K".
>>> COLA arseholes call it choice. I call I ridiculous.
> To each his own. I find it odd that Snit, DFS and now you as well > consider the different editions of Ubuntu as different distros. They are > not. They are the same distro. They just offer a different set of > default packages on the installation medium.
> For what it's worth: I don't necessarily think this is a Good Thing™. > Perhaps Ubuntu should just focus on Gnome as its desktop environment, > and be done with the K and X versions. Those who really want KDE or Xfce > can install it afterwards anyway.
Apparently there are enough people that disagree with you for the derivatives to exist.
Installing one DE, then the other wastes a lot of disk space, which some people may not like. I used to do that when I was primarily running WindowMaker. When I switched to KDE, I started just installing the Gnome/ gtk apps I used. That has worked well for me on Suse, and now PCLinuxOS. When using Ubuntu, I stick to Gnome, except for using k3b.
On 2009-11-08, the following emerged from the brain of Snit:
>>> I think we both believe that I can use my *desktop* system to >>> serve many server functions -
>> Not many. Any.
> Well... not without extensive work... and even then likely not as > well as OS X server (I am on OS X), unless I were to pretty much > make it be OS X Server (or the components I need, of course).
See, to you the system is what the vendor decided it to be. To me it isn't.
You don't need to do any work at all to turn regular OSX into a 'server system'. It has apache. It has postfix. Being a Unix-like OS I pretty sure it has an FTP server and SSH by default as well.
None of the above services (of which any one is enough to set the machine up as a server) will perform better or offer more functionality on the server edition of OSX.
>>> We also agree that, with Linux as least, one can modify their own >>> computer heavily and re-purpose it - setting up the system for >>> different use-cases than it was originally intended for.
>>>> You are willing to replace the term 'server OS' with 'server >>>> system'. By that I understand that you deem a 'server OS' to be >>>> the same as a 'server system'. Is that right?
>>> Yes.
>>>> So now we need to establish what a 'server system' is. It is my >>>> opinion that a system becomes a 'server system' from the moment I >>>> give it a server role.
>>> I would say that a server system is one in which the design has >>> been focused on server needs... not just one which is set with a >>> server role.
>> Setting a system with a server role *is* focusing its 'design' on >> server needs. You are trying to twist it as if it isn't, but it is.
> If I turn on web sharing on my computer I am not changing its > design. I do not see what you are trying to say.
'Web sharing'? Now what's that supposed to be? Another one of those terms that can mean anthing and nothing at the same time. Is this a web server? An FTP server? CIFS? NFS?
If you keep injecting meaningless terms, this discussion is going nowhere (not that it has been going anywhere at all of course).
There is no difference between serving on a lan and serving on a wan. It's just the scope that's different.
You still seem to think that the design is enforced by the vendor. You are free to accept that as a fact. I don't.
>>>> Any server role. Web server (e.g. apache or iis). Mail server >>>> (e.g. postfix or exchange). Directory service (e.g. nis or ad). >>>> File server (e.g. cifs or nfs). Etc. Any one of those, or even >>>> any combination of those. This is OS agnostic. I could do this >>>> with OSX Server, regular OSX, FreeBSD, Debian, Ubuntu Server, >>>> Kubuntu, whatever. Do you still agree with me at this point?
>>> Not quite... as explained above. I think as you modify a system >>> you can get to gray areas as to what it is "designed" for... >>> really, a system need not really be "designed" for either... or >>> could be designed for a bit of both.
>> I'm not interested in 'gray areas'. Why do there have to be 'gray >> areas'?
> There are in many areas of thought. Not all, mind you... but many. > Life is not black and white.
Of course not.
>> I'm talking about *any* machine, setting it up as a server (doesn't >> matter what kind of server, or with how many services) and using it >> for no more than that.
>> Would you disagree I turn a system into a 'server system' by doing >> that?
> You are using the system as a server... I cannot tell what it was > designed for *in general* just by knowing, for example, that it has > a print server on it!
I can tell you. I 'designed' it to be a server. To you 'designing' a system to be a server is a series of hard and work-intensive task you have yet to describe (other than match the vendor's idea of a 'server system'). To me it is enough to slam apache on it and declare it a server.
>> Remember, there's no grey area. The machine will even be tucked >> away in a server room, doors locked, no monitor, keyboard or mouse >> attached. The only cables attached to the machine will be the power >> cord and the network cable.
> To quote below, it might be a "lightweight" server system. Again... > gray areas.
It is a 'server system'. Period. The 'weight' doesn't matter. And even if the 'weight' did matter, you just accepted the fact that I can turn XP into a 'server system', ie. 'server OS'.
>>>> Now, taking all of the above in consideration, why would you say >>>> that installing apache on Windows XP doesn't turn Windows XP into >>>> a 'server OS'? After all a 'server OS' is the same as a 'server >>>> system' (by your own saying), and I really can't see how a server >>>> differs from a server system, other than the term server would >>>> refer to the entire machine rather than just the software running >>>> on that machine.
>>> I think this works well... though they use the term "network >>> operating system":
>> Why don't you just directly answer my question instead of digging >> up some articles from the internet?
> They not only answer the question well, they show that my usage of > language is the common one, or at least real close to it.
>> The basis for this silly discussion is your claim that I couldn't >> possibly turn XP into a 'server OS'.
> What? When did I say that? I noted merely installing Apache on a > desktop OS does not turn it into a server OS... and this is > supported by my desktop OS having Apache *by default* and yet not > being a server OS (as OS X Server is).
So tell me: what do I need to do in your opinion to promote XP to a server OS? I mean, you seem to think I can after all, so you must have a pretty good idea of what I need to do.
At what point becomes XP a 'server OS'?
>> In the meantime you have agreed to replace this term with 'server >> system', so your opinion is that I can't turn XP into a 'server >> system' (yet I can with Ubuntu desktop edition apparently). So >> unless you can give me a perfectly good explanation why a 'server >> system' is different from a 'server', your abovementioned claim >> holds no ground.
> I have explained the difference... and pointed to multiple > references that support my view. You have poo-pooed my ideas and > the references. Not sure what else I can offer - you have your > opinion which you are not going to change. I can accept that. You > have agreed, however, that your use of terms is non-standard.
Well, 'server system' worked fine for me as well. Or am I now using a non-standard term in a non-standard way. Now that is a bit hard to argue against :-p
>> The only possible reason I can think of is that it is your opinion >> that a system can only be a 'server system/OS/whatever' because it >> was sold as one.
> Designed / put together as one.
By anyone then,including the end user?
>> And that, frankly, is a silly claim.
> Well, happily, your straw man is not my claim. :)
I didn't say it was. I said it was the only reason left I could think of. Apparently this isn't, so no I'm really starting to wonder what is...
>>> An operating system that is designed for network use. >>> Normally, it is a complete operating system with file, task >>> and job management; however, with some earlier products, it >>> was a separate component that ran under the OS; for example, >>> LAN Server required OS/2, and LANtastic required DOS.
>>> Unix, Linux, NetWare, Windows 2000 Server and Windows Server >>> 2003 are examples of network operating systems designed for >>> use in stand-alone servers. Such products may also include a >>> Web server, directory services, messaging system, network >>> management and multiprotocol routing capabilities.
>>> Multiuser File Sharing A network operating system (NOS) >>> manages concurrent requests from clients and provides the >>> security necessary in a multiuser environment. A file sharing >>> component is installed in each client machine that interacts >>> with the server to share files and applications as well as >>> network devices such as printers, faxes and modems.
>>> Peer-to-Peer Operating Systems In a peer-to-peer network, such >>> as Windows 98 or XP, where each client machine can also be a >>> server, the OS might still be considered a network operating >>> system, but it is more lightweight than a full-blown NOS. See >>> LAN. -----
>> This is a load of bollocks. Shame on the writer.
> And you disagree with it ... why? Where?
It is just poorly written. In the first paragraph he literally says that a 'network operating system' (yet another meaningless term - all major operating systems are networked nowadays) used to be a component that ran under an OS. Very poor writing style.
In the second paragraph he declared the whole of Unix and Linux as 'network operating systems'. Now how silly is that?
Paragraph three describes about any contemporary networked operating system.
And finally the fourth paragraph calls Windows 98 and Windows XP peer-to-peer networks. Again poor writing style of course. He also seems to think that you can't deploy e.g. XP as a centralized server, which is of course utter bullshit.
>>> Also this: <http://en.wikipedia.org/wiki/Network_Operating_System> >>> ----- A networking operating system is an operating system that >>> contains components and programs that allow a computer on a >>> network to serve requests from other computer for data and provide >>> access to other resources such as printer and file systems.
>>> Features * Add, remove and manage users who wish to use >>> resources on the network. * Allow users to access to the data >>> on the network. This data commonly resides on the server. * >>> Allow users to
>>>>> Different editions, different distros, different default packages, >>>>> different download sites, etc.
>>>> And then add the "K".
>>>> COLA arseholes call it choice. I call I ridiculous.
>> To each his own. I find it odd that Snit, DFS and now you as well >> consider the different editions of Ubuntu as different distros. They are >> not. They are the same distro. They just offer a different set of >> default packages on the installation medium.
>> For what it's worth: I don't necessarily think this is a Good Thing™. >> Perhaps Ubuntu should just focus on Gnome as its desktop environment, >> and be done with the K and X versions. Those who really want KDE or Xfce >> can install it afterwards anyway.
> Apparently there are enough people that disagree with you for the > derivatives to exist.
> Installing one DE, then the other wastes a lot of disk space, which some > people may not like. I used to do that when I was primarily running > WindowMaker. When I switched to KDE, I started just installing the Gnome/ > gtk apps I used. That has worked well for me on Suse, and now PCLinuxOS. > When using Ubuntu, I stick to Gnome, except for using k3b.
> Choice ... :-)
Of course. And if Canonical wished to have a bunch of different editions of their OS, then that's their choice.
I personally like the Debian way much better. There's only one edition (although there are different media to install from, cfr. the netinstall images, the businesscard images, the CD images and the DVD images), and how it is set-up is completely up to the user. Similar with FreeBSD. There is no 'desktop edition', no 'server edition' or no 'media center editions'. It very much hangs together with my idea that an OS can be anything the user wants it to be.
>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>> vice versa in no time.
>> The only thing you have to do is start X, then you have a server. >> Don't know if ssh is enabled by default in linux distros but ssh is a >> server.
> SSH is enabled, even on laptop installs. Both client and server.
This is Ubuntu you're talking about?
Not a good idea in my opinion. This way the machine is immediately suspectible to brute force ssh attacks when directly connected to the internet.
Here's an excerpt of auth.log of a FreeBSD machine I recently set-up as an SSH host:
Nov 5 19:21:16 vortex sshd[988]: Invalid user postgres from 80.221.53.1 Nov 5 19:21:17 vortex sshd[990]: Invalid user administracion from 80.221.53.1 Nov 5 19:21:18 vortex sshd[992]: Invalid user soporte from 80.221.53.1 Nov 5 19:21:19 vortex sshd[994]: Invalid user soporte from 80.221.53.1 Nov 5 19:21:21 vortex sshd[996]: Invalid user postgres from 80.221.53.1 Nov 5 19:21:22 vortex sshd[998]: Invalid user admin from 80.221.53.1 Nov 5 19:21:23 vortex sshd[1000]: Invalid user admin from 80.221.53.1 Nov 5 19:21:24 vortex sshd[1002]: Invalid user test from 80.221.53.1 Nov 5 19:21:28 vortex sshd[1008]: Invalid user support from 80.221.53.1 Nov 5 19:21:29 vortex sshd[1010]: Invalid user support from 80.221.53.1 Nov 5 19:21:31 vortex sshd[1012]: Invalid user andrea from 80.221.53.1 Nov 5 19:21:32 vortex sshd[1014]: Invalid user andrea from 80.221.53.1 Nov 5 19:21:33 vortex sshd[1016]: Invalid user ubuntu from 80.221.53.1 Nov 5 19:21:34 vortex sshd[1018]: Invalid user ubuntu from 80.221.53.1 Nov 5 19:21:38 vortex sshd[1024]: Invalid user prova from 80.221.53.1 Nov 5 19:21:40 vortex sshd[1026]: Invalid user prova from 80.221.53.1 Nov 5 19:21:41 vortex sshd[1028]: Invalid user nagios from 80.221.53.1 Nov 5 19:21:42 vortex sshd[1030]: Invalid user nagios from 80.221.53.1 Nov 5 19:21:45 vortex sshd[1034]: Invalid user vladimir from 80.221.53.1 Nov 5 19:21:47 vortex sshd[1036]: Invalid user vladimir from 80.221.53.1 Nov 5 19:21:48 vortex sshd[1038]: Invalid user sales from 80.221.53.1 Nov 5 19:21:50 vortex sshd[1040]: Invalid user sales from 80.221.53.1 Nov 5 19:21:51 vortex sshd[1042]: Invalid user service from 80.221.53.1 Nov 5 19:21:53 vortex sshd[1044]: Invalid user service from 80.221.53.1 Nov 5 19:21:59 vortex sshd[1052]: Invalid user test from 80.221.53.1 Nov 5 19:22:00 vortex sshd[1054]: Invalid user user from 80.221.53.1 Nov 5 19:22:02 vortex sshd[1069]: Invalid user alex from 80.221.53.1 Nov 5 19:22:05 vortex sshd[1073]: Invalid user test from 80.221.53.1 Nov 5 19:22:07 vortex sshd[1075]: Invalid user iser from 80.221.53.1 Nov 5 19:22:08 vortex sshd[1077]: Invalid user patrick from 80.221.53.1 Nov 5 19:22:10 vortex sshd[1079]: Invalid user patrick from 80.221.53.1 Nov 5 19:22:11 vortex sshd[1081]: Invalid user patrick from 80.221.53.1 Nov 5 19:22:13 vortex sshd[1083]: Invalid user patrick from 80.221.53.1 Nov 5 19:22:15 vortex sshd[1085]: Invalid user testing from 80.221.53.1 Nov 5 19:22:16 vortex sshd[1087]: Invalid user renato from 80.221.53.1 Nov 5 19:22:23 vortex sshd[1095]: Invalid user test from 80.221.53.1 Nov 5 19:22:25 vortex sshd[1099]: Invalid user web from 80.221.53.1 Nov 5 19:22:28 vortex sshd[1103]: Invalid user test from 80.221.53.1 Nov 5 19:22:32 vortex sshd[1107]: Invalid user test from 80.221.53.1 Nov 5 19:22:35 vortex sshd[1111]: Invalid user admin from 80.221.53.1 Nov 5 19:22:37 vortex sshd[1113]: Invalid user cmc from 80.221.53.1 Nov 5 19:22:40 vortex sshd[1117]: Invalid user spam from 80.221.53.1 Nov 5 19:22:42 vortex sshd[1119]: Invalid user cafe from 80.221.53.1 Nov 5 19:22:46 vortex sshd[1125]: Invalid user master from 80.221.53.1 Nov 5 19:22:48 vortex sshd[1127]: Invalid user master from 80.221.53.1 Nov 5 19:22:51 vortex sshd[1131]: Invalid user oracle from 80.221.53.1 Nov 5 19:22:54 vortex sshd[1135]: Invalid user oracle from 80.221.53.1 Nov 5 19:23:18 vortex sshd[1163]: Invalid user average from 80.221.53.1 Nov 5 19:23:21 vortex sshd[1167]: Invalid user paulo from 80.221.53.1 Nov 5 19:23:23 vortex sshd[1169]: Invalid user eric from 80.221.53.1 Nov 5 19:23:27 vortex sshd[1173]: Invalid user camera from 80.221.53.1 Nov 5 19:23:29 vortex sshd[1175]: Invalid user rubens from 80.221.53.1 Nov 5 19:23:31 vortex sshd[1177]: Invalid user squid from 80.221.53.1 Nov 5 19:23:33 vortex sshd[1179]: Invalid user squid from 80.221.53.1 Nov 5 19:23:35 vortex sshd[1181]: Invalid user paolo from 80.221.53.1 Nov 5 19:23:37 vortex sshd[1183]: Invalid user paolo from 80.221.53.1 Nov 5 19:23:40 vortex sshd[1185]: Invalid user charles from 80.221.53.1 Nov 5 19:23:42 vortex sshd[1187]: Invalid user charles from 80.221.53.1 Nov 5 19:23:44 vortex sshd[1189]: Invalid user spammail from 80.221.53.1
This started immediately after I started the sshd daemon for the first time.
Of course I changed the port number and replaced pam authentication with pubkey authentication. Now auth.log is golden silence most of the time.
-- Let him choose out of my files, his projects to accomplish. -- Shakespeare, "Coriolanus"
>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>> vice versa in no time.
>>> The only thing you have to do is start X, then you have a server. >>> Don't know if ssh is enabled by default in linux distros but ssh is a >>> server.
>> SSH is enabled, even on laptop installs. Both client and server.
> This is Ubuntu you're talking about?
No, Debian. If you select a Desktop setup.
> Not a good idea in my opinion. This way the machine is immediately > suspectible to brute force ssh attacks when directly connected to the > internet.
No shit. My logs match yours! The funny ones are the ones that try root directly.
> Here's an excerpt of auth.log of a FreeBSD machine I recently set-up > as an SSH host:
> Nov 5 19:21:16 vortex sshd[988]: Invalid user postgres from 80.221.53.1 > Nov 5 19:21:17 vortex sshd[990]: Invalid user administracion from 80.221.53.1
> . . .
> This started immediately after I started the sshd daemon for the first > time.
> Of course I changed the port number and replaced pam authentication > with pubkey authentication. Now auth.log is golden silence most of the > time.
I can't change the port number; my workplace blocks most of the non-standard ports.
TomB <tommy.bongae...@gmail.com> writes: > On 2009-11-08, the following emerged from the brain of Chris Ahlstrom: >> Gunnar Gren pulled this Usenet boner:
>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>> vice versa in no time.
>>> The only thing you have to do is start X, then you have a server. >>> Don't know if ssh is enabled by default in linux distros but ssh is a >>> server.
>> SSH is enabled, even on laptop installs. Both client and server.
> This is Ubuntu you're talking about?
> Not a good idea in my opinion. This way the machine is immediately > suspectible to brute force ssh attacks when directly connected to the > internet.
Configure good keys and a password to go with them and you're safe as houses when you couple it with something like DenyHosts.
On 2009-11-08, the following emerged from the brain of Chris Ahlstrom:
> TomB pulled this Usenet boner:
>> On 2009-11-08, the following emerged from the brain of Chris Ahlstrom: >>> SSH is enabled, even on laptop installs. Both client and server.
>> This is Ubuntu you're talking about?
> No, Debian. If you select a Desktop setup.
Shame on Debian then.
>> Not a good idea in my opinion. This way the machine is immediately >> suspectible to brute force ssh attacks when directly connected to the >> internet.
> No shit. My logs match yours! The funny ones are the ones that try root > directly.
I understand that Debian is sensible enough then to disable root login over ssh then. Well, that's something.
But still, I can image people actually using credentials such as frank/frank123 and give root a very simple password too. When such a machine happens to be directly connected to the internet, it will be a matter of hours - perhaps even minutes - before it is compromised.
>> Here's an excerpt of auth.log of a FreeBSD machine I recently set-up >> as an SSH host:
>> Nov 5 19:21:16 vortex sshd[988]: Invalid user postgres from 80.221.53.1 >> Nov 5 19:21:17 vortex sshd[990]: Invalid user administracion from 80.221.53.1
>> . . .
>> This started immediately after I started the sshd daemon for the first >> time.
>> Of course I changed the port number and replaced pam authentication >> with pubkey authentication. Now auth.log is golden silence most of the >> time.
> I can't change the port number; my workplace blocks most of the non-standard > ports.
If your password is strong enough (or if you use pubkey authentication), you should be quite safe anyway.
Shit, I'm still amazed Debian enables SSH by default on a full desktop install...
>>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>>> vice versa in no time.
>>>> The only thing you have to do is start X, then you have a server. >>>> Don't know if ssh is enabled by default in linux distros but ssh is a >>>> server.
>>> SSH is enabled, even on laptop installs. Both client and server.
>> This is Ubuntu you're talking about?
>> Not a good idea in my opinion. This way the machine is immediately >> suspectible to brute force ssh attacks when directly connected to the >> internet.
> Configure good keys and a password to go with them and you're safe as > houses when you couple it with something like DenyHosts.
> On 2009-11-08, the following emerged from the brain of Chris Ahlstrom:
>> I can't change the port number; my workplace blocks most of the non-standard >> ports.
> If your password is strong enough (or if you use pubkey > authentication), you should be quite safe anyway.
I make all my passwords strong, and I have a different password for every system. This is at home; at work, I have no choice but to make strong passwords.
I even force my wife to make better passwords on her Billy Box.
> Shit, I'm still amazed Debian enables SSH by default on a full desktop > install...
> Thanks for the link. Looks like a very useful additional layer of > security. I may actually start to use this.
I am surprised you don't know about this, Tom. It's been around ever since I started using Linux, nearly a decade (RedHat). It's part of tcp-wrappers. You don't need that site, just set up xinetd; it applies to more than SSH.
(Disclaimer: it's been a long time since I used tcp wrappers. Way back before I was behind a hardware firewall).
The denyhosts package is part of Debian.
-- O, what a tangled web we weave, When first we practice to deceive. -- Sir Walter Scott, "Marmion"
>>>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>>>> vice versa in no time.
>>>>> The only thing you have to do is start X, then you have a server. >>>>> Don't know if ssh is enabled by default in linux distros but ssh is a >>>>> server.
>>>> SSH is enabled, even on laptop installs. Both client and server.
>>> This is Ubuntu you're talking about?
>>> Not a good idea in my opinion. This way the machine is immediately >>> suspectible to brute force ssh attacks when directly connected to the >>> internet.
>> Configure good keys and a password to go with them and you're safe as >> houses when you couple it with something like DenyHosts.
> Thanks for the link. Looks like a very useful additional layer of > security. I may actually start to use this.
Seriously, default ssh is fine. Use public key encryption and then you can ssh in with NO password from your trusted machines but configure the keys to ALSO need a password from other machines.
In article <Z-2dnYGZ393jNmvXnZ2dnUVZ_qRi4...@supernews.com>,
Rick <n...@mail.invalid> wrote: > > For what it's worth: I don't necessarily think this is a Good Thing™. > > Perhaps Ubuntu should just focus on Gnome as its desktop environment, > > and be done with the K and X versions. Those who really want KDE or Xfce > > can install it afterwards anyway.
> Apparently there are enough people that disagree with you for the > derivatives to exist.
> Installing one DE, then the other wastes a lot of disk space, which some > people may not like. I used to do that when I was primarily running
If only they used some sort of packaging system that allowed for uninstalling packages, so that you could switch desktop environments by installing the new one, and then uninstalling the old one....
>> Thanks for the link. Looks like a very useful additional layer of >> security. I may actually start to use this.
> I am surprised you don't know about this, Tom. It's been around > ever since I started using Linux, nearly a decade (RedHat). It's > part of tcp-wrappers. You don't need that site, just set up xinetd; > it applies to more than SSH.
I know (x)inetd and tcpwrappers well. I just didn't know of this script that automatically adds entries in hosts.denied based on what it finds in the logs. This is in Debian you say? Never noticed it; never heard of it.
Ah, even 10 years isn't enough :-p
> (Disclaimer: it's been a long time since I used tcp wrappers. Way > back before I was behind a hardware firewall).
Hardware firewalls and packet filters indeed have largely supplanted tcpwrappers. Haven't used them for a while as well.
> The denyhosts package is part of Debian.
Indeed it is. Nice. Thanks too.
So this thread turns out to be usefull after all ;-)
-- An exotic journey in downtown Newark is in your future.
On 2009-11-08, the following emerged from the brain of Gunnar Gren:
> Den 2009-11-08 skrev TomB <tommy.bongae...@gmail.com>: >> On 2009-11-08, the following emerged from the brain of Gunnar Gren: >>> 2009-11-04 TomB <tommy.bongae...@gmail.com>:
>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>> vice versa in no time.
>>> The only thing you have to do is start X, then you have a server.
>> At least Debian starts X with the --nolisten tcp flag as default, so >> that's not exactly true.
> FreeBSD does not. X is a server program and so is cups, lpd and almost > every such program that runs on unix. But I use it as desktop and server..
I don't use X on my FreeBSD machines, so I wouldn't know.
>>> Don't know if ssh is enabled by default in linux distros but ssh is >>> a server.
>> It is. I have a server at work that acts as an ssh proxy for remote >> administration of the other servers on the network. It's a FreeBSD >> virtual machine doing nothing else than listening for incoming ssh >> connections (pubkey authentication only of course). The entire >> installation takes less than 300MB and the only running service is >> sshd. I'm sure Snit wouldn't think of this as a server >> system/OS/whatever, but it is. I know it, because I set it up to be >> one.
>>>> I can turn an Ubuntu desktop installation into an Ubuntu server and >>>> vice versa in no time.
>>> The only thing you have to do is start X, then you have a server. >>> Don't know if ssh is enabled by default in linux distros but ssh is a >>> server.
>> SSH is enabled, even on laptop installs. Both client and server.
> Aha. Ok. I haven't used linux for many years now, the last distro I > managed was trustix 1.x and in FreeBSD you have to enable ssh in rc.conf.
The only sensible way imho. Make the user work hard to shoot himself in the foot.
-- Q: Why did the lone ranger kill Tonto? A: He found out what "kimosabe" really means.
TomB stated in post 20091108122237....@usenet.drumscum.be on 11/8/09 6:36 AM:
> On 2009-11-08, the following emerged from the brain of Snit:
>>>> I think we both believe that I can use my *desktop* system to >>>> serve many server functions -
>>> Not many. Any.
>> Well... not without extensive work... and even then likely not as >> well as OS X server (I am on OS X), unless I were to pretty much >> make it be OS X Server (or the components I need, of course).
> See, to you the system is what the vendor decided it to be. To me it > isn't.
I would hope the vender was able to get the system to be what the decided it should be! With Linux, though, they are limited... for example, Ubuntu clearly is made to be as consistent as reasonably possible but still falls short in many ways.
> You don't need to do any work at all to turn regular OSX into a > 'server system'. It has apache. It has postfix. Being a Unix-like OS I > pretty sure it has an FTP server and SSH by default as well.
> None of the above services (of which any one is enough to set the > machine up as a server) will perform better or offer more > functionality on the server edition of OSX.
They have a better UI for getting them set up. The system, after all, was *designed* to be a server system.
>>>> We also agree that, with Linux as least, one can modify their own >>>> computer heavily and re-purpose it - setting up the system for >>>> different use-cases than it was originally intended for.
> Left this bit in intentionally.
Right... but above you imply I do not accept this. Why?
>>>>> You are willing to replace the term 'server OS' with 'server >>>>> system'. By that I understand that you deem a 'server OS' to be >>>>> the same as a 'server system'. Is that right?
>>>> Yes.
>>>>> So now we need to establish what a 'server system' is. It is my >>>>> opinion that a system becomes a 'server system' from the moment I >>>>> give it a server role.
>>>> I would say that a server system is one in which the design has >>>> been focused on server needs... not just one which is set with a >>>> server role.
>>> Setting a system with a server role *is* focusing its 'design' on >>> server needs. You are trying to twist it as if it isn't, but it is.
>> If I turn on web sharing on my computer I am not changing its >> design. I do not see what you are trying to say.
> 'Web sharing'? Now what's that supposed to be? Another one of those > terms that can mean anthing and nothing at the same time. Is this a > web server? An FTP server? CIFS? NFS?
Is web sharing the same as FTP? Huh? Web sharing is, well, sharing things via a web server. Not sure where the confusion came from there.
> If you keep injecting meaningless terms, this discussion is going > nowhere (not that it has been going anywhere at all of course).
Now "web sharing" is a meaningless term to you. OK. Then OS X Desktop has "meaningless" web server capabilities... right? At least in the GUI.
> There is no difference between serving on a lan and serving on a wan. > It's just the scope that's different.
> You still seem to think that the design is enforced by the vendor. You > are free to accept that as a fact. I don't.
Enforced? What? Remember:
We also agree that, with Linux as least, one can modify their own computer heavily and re-purpose it - setting up the system for different use-cases than it was originally intended for.
>>>>> Any server role. Web server (e.g. apache or iis). Mail server >>>>> (e.g. postfix or exchange). Directory service (e.g. nis or ad). >>>>> File server (e.g. cifs or nfs). Etc. Any one of those, or even >>>>> any combination of those. This is OS agnostic. I could do this >>>>> with OSX Server, regular OSX, FreeBSD, Debian, Ubuntu Server, >>>>> Kubuntu, whatever. Do you still agree with me at this point?
>>>> Not quite... as explained above. I think as you modify a system >>>> you can get to gray areas as to what it is "designed" for... >>>> really, a system need not really be "designed" for either... or >>>> could be designed for a bit of both.
>>> I'm not interested in 'gray areas'. Why do there have to be 'gray >>> areas'?
>> There are in many areas of thought. Not all, mind you... but many. >> Life is not black and white.
> Of course not.
Then why argue with me so much over the concept of gray areas and pretend like to even mention them somehow is a bad thing?
>>> I'm talking about *any* machine, setting it up as a server (doesn't >>> matter what kind of server, or with how many services) and using it >>> for no more than that.
>>> Would you disagree I turn a system into a 'server system' by doing >>> that?
>> You are using the system as a server... I cannot tell what it was >> designed for *in general* just by knowing, for example, that it has >> a print server on it!
> I can tell you. I 'designed' it to be a server. To you 'designing' a > system to be a server is a series of hard and work-intensive task you > have yet to describe (other than match the vendor's idea of a 'server > system'). To me it is enough to slam apache on it and declare it a > server.
I have described and pointed you to external resources. You poo-pood them and me... and never backed up your view that any system with any server capabilities magically becomes a server system. Your only attempt is to re-defined the concept of a server system. Semantics.
>>> Remember, there's no grey area. The machine will even be tucked >>> away in a server room, doors locked, no monitor, keyboard or mouse >>> attached. The only cables attached to the machine will be the power >>> cord and the network cable.
>> To quote below, it might be a "lightweight" server system. Again... >> gray areas.
> It is a 'server system'. Period. The 'weight' doesn't matter. And even > if the 'weight' did matter, you just accepted the fact that I can turn > XP into a 'server system', ie. 'server OS'.
And I can turn a large piece of cardboard and 20 dogs on chains into "mass transit". Not exactly the Chicago el. Again, you are playing semantics.
>>>>> Now, taking all of the above in consideration, why would you say >>>>> that installing apache on Windows XP doesn't turn Windows XP into >>>>> a 'server OS'? After all a 'server OS' is the same as a 'server >>>>> system' (by your own saying), and I really can't see how a server >>>>> differs from a server system, other than the term server would >>>>> refer to the entire machine rather than just the software running >>>>> on that machine.
>>>> I think this works well... though they use the term "network >>>> operating system":
>>> Why don't you just directly answer my question instead of digging >>> up some articles from the internet?
>> They not only answer the question well, they show that my usage of >> language is the common one, or at least real close to it.
>>> The basis for this silly discussion is your claim that I couldn't >>> possibly turn XP into a 'server OS'.
>> What? When did I say that? I noted merely installing Apache on a >> desktop OS does not turn it into a server OS... and this is >> supported by my desktop OS having Apache *by default* and yet not >> being a server OS (as OS X Server is).
> So tell me: what do I need to do in your opinion to promote XP to a > server OS? I mean, you seem to think I can after all, so you must have > a pretty good idea of what I need to do.
> At what point becomes XP a 'server OS'?
Now you are looking for an exact formula when I have noted, repeatedly, that there are gray areas. To see what Windows looks like as a server OS, look at Windows server... that should give you a basic idea.
>>> In the meantime you have agreed to replace this term with 'server >>> system', so your opinion is that I can't turn XP into a 'server >>> system' (yet I can with Ubuntu desktop edition apparently). So >>> unless you can give me a perfectly good explanation why a 'server >>> system' is different from a 'server', your abovementioned claim >>> holds no ground.
>> I have explained the difference... and pointed to multiple >> references that support my view. You have poo-pooed my ideas and >> the references. Not sure what else I can offer - you have your >> opinion which you are not going to change. I can accept that. You >> have agreed, however, that your use of terms is non-standard.
> Well, 'server system' worked fine for me as well. Or am I now using a > non-standard term in a non-standard way. Now that is a bit hard to > argue against :-p
>>> The only possible reason I can think of is that it is your opinion >>> that a system can only be a 'server system/OS/whatever' because it >>> was sold as one.
>>>> An operating system that is designed for network use. >>>> Normally, it is a complete operating system with file, task >>>> and job management; however, with some earlier products, it >>>> was a separate component that ran under the OS; for example, >>>> LAN Server required OS/2, and LANtastic required DOS.
>>>> Unix, Linux, NetWare, Windows 2000 Server and Windows Server >>>> 2003 are examples of network operating systems designed for >>>> use in stand-alone servers. Such products may also include a >>>> Web server, directory services, messaging system, network >>>> management and multiprotocol routing capabilities.
TomB stated in post 20091108115241....@usenet.drumscum.be on 11/8/09 4:11 AM:
>> This whole thing came from me noting that some systems are designed to be >> servers and some are designed to be desktops... and the difference is far >> more than if Apache is or is not installed.
> As I said numerous times: it doesn't matter how many server > applications are running to turn a system into a 'server system'. > That's a vague distinction (how many are needed?) that you try to > force on the notion of what exacltly makes a system a 'server system', > but doesn't make any sense at all.
While you might not be able to make sense of it, recall I have pointed you to multiple sources which show they *do* understand the concept. One even discussed your specific misunderstanding.
> It only makes remotely sense if you are saying that the difference > between a 'server OS' and a 'desktop OS' is defined by the > distributor, and that the end user can't possibly setup one as the > other.
Not sure why you ignore this:
We also agree that, with Linux as least, one can modify their own computer heavily and re-purpose it - setting up the system for different use-cases than it was originally intended for.
On 2009-11-08, the following emerged from the brain of Snit:
>>>>> I think we both believe that I can use my *desktop* system to >>>>> serve many server functions -
>>>> Not many. Any.
>>> Well... not without extensive work... and even then likely not as >>> well as OS X server (I am on OS X), unless I were to pretty much >>> make it be OS X Server (or the components I need, of course).
>> See, to you the system is what the vendor decided it to be. To me >> it isn't.
> I would hope the vender was able to get the system to be what the > decided it should be!
Uh? I never said anything about the vendor being able to do whatever.
>> You don't need to do any work at all to turn regular OSX into a >> 'server system'. It has apache. It has postfix. Being a Unix-like >> OS I pretty sure it has an FTP server and SSH by default as well.
>> None of the above services (of which any one is enough to set the >> machine up as a server) will perform better or offer more >> functionality on the server edition of OSX.
> They have a better UI for getting them set up. The system, after > all, was *designed* to be a server system.
So you feel that it is deciding factor to have a graphical configuration interface to the server software? Now *that* is an uncommon opinion. You're basically saying that all those headless server machines out there in the wild can't possible be 'server systems'.
>>>>> We also agree that, with Linux as least, one can modify their >>>>> own computer heavily and re-purpose it - setting up the system >>>>> for different use-cases than it was originally intended for.
>> Left this bit in intentionally.
> Right... but above you imply I do not accept this. Why?
I don't imply that you don't accept this. I'm noting that with heavy modification you actually mean turning the system into the vendor's idea of a server.
>> 'Web sharing'? Now what's that supposed to be? Another one of those >> terms that can mean anthing and nothing at the same time. Is this a >> web server? An FTP server? CIFS? NFS?
> Is web sharing the same as FTP? Huh? Web sharing is, well, sharing > things via a web server. Not sure where the confusion came from > there.
Well, an FTP server facing the internet at large is 'web sharing' too.
(Come to think of it: in fact the term 'web server' is meaningless too. Never thought about that. Http server would be a more accurate description.)
>> If you keep injecting meaningless terms, this discussion is going >> nowhere (not that it has been going anywhere at all of course).
> Now "web sharing" is a meaningless term to you. OK. Then OS X > Desktop has "meaningless" web server capabilities... right? At > least in the GUI.
Well the common term for this kind of serving is 'web serving'. Why use yet another term?
>> There is no difference between serving on a lan and serving on a >> wan. It's just the scope that's different.
>> You still seem to think that the design is enforced by the vendor. >> You are free to accept that as a fact. I don't.
> Enforced? What? Remember:
> We also agree that, with Linux as least, one can modify their > own computer heavily and re-purpose it - setting up the system > for different use-cases than it was originally intended for.
See above.
>>>> I'm not interested in 'gray areas'. Why do there have to be 'gray >>>> areas'?
>>> There are in many areas of thought. Not all, mind you... but >>> many. Life is not black and white.
>> Of course not.
> Then why argue with me so much over the concept of gray areas and > pretend like to even mention them somehow is a bad thing?
I'm not saying gray areas are a bad thing. There are many inbetweens indeed. But when I set-up a machine as a server and nothing but a server there is no gray area. It's a server, no matter what OS I'm using, or how many server programs I am running on it.
>> So tell me: what do I need to do in your opinion to promote XP to a >> server OS? I mean, you seem to think I can after all, so you must >> have a pretty good idea of what I need to do.
>> At what point becomes XP a 'server OS'?
> Now you are looking for an exact formula when I have noted, > repeatedly, that there are gray areas. To see what Windows looks > like as a server OS, look at Windows server... that should give you > a basic idea.
Another example of you thinking that a 'server OS/system/whatever' is what the vendor dictates it to be.
>> It is just poorly written. In the first paragraph he literally says >> that a 'network operating system' (yet another meaningless term - >> all major operating systems are networked nowadays) used to be a >> component that ran under an OS. Very poor writing style.
> See how you go out of your way to push the idea that any networked > system is a network operating system.
But they are. Servers and clients alike. I'm not pushing anything.
> Seriously, you should take some classes on this or work in an > environment where people use these terms.
Well, I did take classes, and I do work in such an environment.
>> In the second paragraph he declared the whole of Unix and Linux as >> 'network operating systems'. Now how silly is that?
>> Paragraph three describes about any contemporary networked >> operating system.
>> And finally the fourth paragraph calls Windows 98 and Windows XP >> peer-to-peer networks. Again poor writing style of course. He also >> seems to think that you can't deploy e.g. XP as a centralized >> server, which is of course utter bullshit.
> Your main reason for saying it is poor writing is that he is using > standard terms. Odd.
No, he is using all kinds of terms willy nilly. There isn't a single coherent thought in the entire article.
>>>>> Also this: >>>>> <http://en.wikipedia.org/wiki/Network_Operating_System> ----- A >>>>> networking operating system is an operating system that contains >>>>> components and programs that allow a computer on a network to >>>>> serve requests from other computer for data and provide access >>>>> to other resources such as printer and file systems.
>>>>> Features * Add, remove and manage users who wish to use >>>>> resources on the network. * Allow users to access to the >>>>> data on the network. This data commonly resides on the >>>>> server. * Allow users to access data found on other network >>>>> such as the internet. * Allow users to access hardware >>>>> connected to the network. * Protect data and services >>>>> located on the network.
>>>>> Network operating system features may include: * basic >>>>> support for hardware ports * security features such as >>>>> authentication, authorization, login restrictions, and >>>>> access control * name services and directory services * >>>>> file, print, data storage, backup and replication services * >>>>> remote access * system management * network administration >>>>> and auditing tools with graphic interfaces * clustering >>>>> capabilities * fault tolerance and high availability
>>>> I can basically agree with the above.
>>> But just installing Apache does not do that!
>> Not all of that no. But that's not a requirement. It's even poor >> sever design to have all those services running if the machine is >> just meant to be e.g. a web server.
> You now want all the services running to show it was designed to do > them well? Huh?
I'll bounce that right back: why have services installed you aren't using anyway?
I'd say I made a better server (system/OS/whatever) if I stripped off all but the bare essentials.
>>>>> Misconception Some device operating systems, including Mac >>>>> OS X and all versions of Microsoft Windows since Windows >>>>> 2000, include NOS features. A NOS is an OS that has been >>>>> specifically written to implement and maintain networks. >>>>> -----
>>>> But not with this. Poorly written too, this last section.
>>> So why disagree? Where is it wrong?
>> It's not clear in what it tries to say. The header is >> 'Misconception'. What is this misconception?
> One very much like what you have been pushing. Kinda funny to find > it, really. :)
>> That some 'device operating systems' (another meaningless term - >> what the hell does he mean by that?) include 'NOS' features? Or is >> the misconception that a 'NOS' is an OS that has been specifically >> written to implement and maintain networks?
>> The first one is actually a fact (you said so yourself), so it must >> be the second one that's the misconception. A 'NOS' is not >> specifically written to implement and maintain networks after all >> then?
>> Now I *am* confused. :-p
>> (feel free to use this against me!)
I still am confused. Which is the misconception then?
-- Grabel's Law: 2 is not equal to 3 -- not even for large values of 2.
> On 2009-11-08, the following emerged from the brain of Snit:
>>>>>> I think we both believe that I can use my *desktop* system to >>>>>> serve many server functions -
>>>>> Not many. Any.
>>>> Well... not without extensive work... and even then likely not as >>>> well as OS X server (I am on OS X), unless I were to pretty much >>>> make it be OS X Server (or the components I need, of course).
>>> See, to you the system is what the vendor decided it to be. To me >>> it isn't.
>> I would hope the vender was able to get the system to be what the >> decided it should be!
> Uh? I never said anything about the vendor being able to do whatever.
You noted that, to me, a system is what a vender decided it should be. Now, sure, there may be times when the vender cannot make what they want or whatever, but, sure, I would hope a vender supplies a system like they wanted to!
>>> You don't need to do any work at all to turn regular OSX into a >>> 'server system'. It has apache. It has postfix. Being a Unix-like >>> OS I pretty sure it has an FTP server and SSH by default as well.
>>> None of the above services (of which any one is enough to set the >>> machine up as a server) will perform better or offer more >>> functionality on the server edition of OSX.
>> They have a better UI for getting them set up. The system, after >> all, was *designed* to be a server system.
> So you feel that it is deciding factor to have a graphical > configuration interface to the server software? Now *that* is an > uncommon opinion. You're basically saying that all those headless > server machines out there in the wild can't possible be 'server > systems'.
Huh? What? Where did you make that jump? I gave an example of why *one* system was clearly designed to be a server... not that all servers must be designed the same way!
At this point you are just trying to twist my words. Remember, the concept here is not complex:
A well designed system should be designed for a purpose. Two major classes of purposes are systems to directly serve the needs of a desktop user and systems designed to serve the needs of networked users. Systems focused on desktop needs are often said to have a "Desktop OS" and systems focused on server / network needs are often said to have a "Server OS" or "Network OS". I have agreed that the terms "Desktop System", "Server System" and "Network System" are more accurate, even if not what is commonly used. I have also agreed that most desktop systems offer some network / server functionality... which is not to say that is the focus of their design. To use an ever-popular vehicle analogy: If I move a load of bricks with my *passenger* car, that does not turn it into a truck. You can, of course, modify a car to become a truck... and there are some vehicles with properties of both passenger vehicles and trucks.
In addition to the above being, to most experienced users, fairly common knowledge, I have pointed you to multiple resources to support this view. Your reaction is just to declare myself and the resources wrong and just repeat your claims that you are right with *no* support. None. Do you have any?
Not that it really matter... as I have noted, this is mostly just semantics. You are using uncommon terminology (something you have agreed you are doing) and just want to call others wrong for not using *your* terminology. You even go so far as to insist that other sources must be using your terminology in their descriptions. That is a bit silly of you.
>>>>>>> I think we both believe that I can use my *desktop* system to >>>>>>> serve many server functions -
>>>>>> Not many. Any.
>>>>> Well... not without extensive work... and even then likely not >>>>> as well as OS X server (I am on OS X), unless I were to pretty >>>>> much make it be OS X Server (or the components I need, of >>>>> course).
>>>> See, to you the system is what the vendor decided it to be. To me >>>> it isn't.
>>> I would hope the vender was able to get the system to be what the >>> decided it should be!
>> Uh? I never said anything about the vendor being able to do >> whatever.
> You noted that, to me, a system is what a vender decided it should > be. Now, sure, there may be times when the vender cannot make what > they want or whatever, but, sure, I would hope a vender supplies a > system like they wanted to!
You're completely missing the point, or you are deliberately avoiding the point.
>>>> You don't need to do any work at all to turn regular OSX into a >>>> 'server system'. It has apache. It has postfix. Being a Unix-like >>>> OS I pretty sure it has an FTP server and SSH by default as well.
>>>> None of the above services (of which any one is enough to set the >>>> machine up as a server) will perform better or offer more >>>> functionality on the server edition of OSX.
>>> They have a better UI for getting them set up. The system, after >>> all, was *designed* to be a server system.
>> So you feel that it is deciding factor to have a graphical >> configuration interface to the server software? Now *that* is an >> uncommon opinion. You're basically saying that all those headless >> server machines out there in the wild can't possible be 'server >> systems'.
> Huh? What? Where did you make that jump? I gave an example of why > *one* system was clearly designed to be a server... not that all > servers must be designed the same way!
> At this point you are just trying to twist my words.
I'm not. I'm just trying to figure out what you think is essential to a 'server OS/system'. Throughout the discussion we have established that it isn't the OS itself, but rather the server software and the functionality it provides. That implies that I can turn e.g. XP into a server OS/system. I gave the example of installing apache (any server packages would've done really), but to you that was not enough. So I an asking you what *is* enough. Since you can say what isn't, you must be able to say what is too. Only, you can't seem to come up with an answer.
> Remember, the concept here is not complex:
Indeed it isn't. When a system serves, it is a server. When it does nothing but serving, it is a dedicated server. It has nothing to do with the OS or the edition of the OS on the machine. The only thing the 'server edition' of an OS does is give you a set of applications that allow you to get started right away, and perhaps some optimizations in e.g. the kernel and system settings.
> In addition to the above being, to most experienced users, fairly > common knowledge, I have pointed you to multiple resources to > support this view. Your reaction is just to declare myself and the > resources wrong and just repeat your claims that you are right with > *no* support. None. Do you have any?
I'm sorry, but you are lying again. I have given support for my claims throughout the entire thread.
> Not that it really matter... as I have noted, this is mostly just > semantics. You are using uncommon terminology (something you have > agreed you are doing) and just want to call others wrong for not > using *your* terminology. You even go so far as to insist that > other sources must be using your terminology in their descriptions. > That is a bit silly of you.
I did never insist on usage of any terms. I've noted that a lot of the terms are meaningless and ambiguous, and they are.
-- Unless you love someone, nothing else makes any sense. -- E.E. Cummings
>> Thanks for the link. Looks like a very useful additional layer of >> security. I may actually start to use this.
> I am surprised you don't know about this, Tom. It's been around ever since > I started using Linux, nearly a decade (RedHat). It's part of tcp-wrappers. > You don't need that site, just set up xinetd; it applies to more than SSH.
I'm rather surprised that he didn't know too. I've used 'xinetd' for about 9/10 years, ever since I started using SuSE Linux.
> (Disclaimer: it's been a long time since I used tcp wrappers. Way back > before I was behind a hardware firewall).
> The denyhosts package is part of Debian.
-- Linux. The Malicious Software Removal tool which wipes Windows from your PC in seconds!
>>> Thanks for the link. Looks like a very useful additional layer of >>> security. I may actually start to use this.
>> I am surprised you don't know about this, Tom. It's been around ever since >> I started using Linux, nearly a decade (RedHat). It's part of tcp-wrappers. >> You don't need that site, just set up xinetd; it applies to more than SSH.
> I'm rather surprised that he didn't know too. > I've used 'xinetd' for about 9/10 years, ever since I started using SuSE > Linux.
Poor Willy.
Has ZERO idea what we are talking about.
No Willy. We are not talking about xinetd itself but rather denyhosts.
Fail2ban is a more powerful solution you might also consider.
>> (Disclaimer: it's been a long time since I used tcp wrappers. Way back >> before I was behind a hardware firewall).
TomB stated in post 20091108235505....@usenet.drumscum.be on 11/8/09 5:03 PM:
...
>>> Uh? I never said anything about the vendor being able to do >>> whatever.
>> You noted that, to me, a system is what a vender decided it should >> be. Now, sure, there may be times when the vender cannot make what >> they want or whatever, but, sure, I would hope a vender supplies a >> system like they wanted to!
> You're completely missing the point, or you are deliberately avoiding > the point.
I am *making* a point. You claimed:
See, to you the system is what the vendor decided it to be.
And I am noting it is best if a system is what the vender decided it to be, at least when it is shipped / downloaded. After that, of course, especially with Linux, a user can make whatever modifications they like. With OS X and Windows things are not always so easy. This is a benefit of Linux (though you and I have agreed it is possible to re-purpose systems built by Apple and MS as well).
>>> So you feel that it is deciding factor to have a graphical >>> configuration interface to the server software? Now *that* is an >>> uncommon opinion. You're basically saying that all those headless >>> server machines out there in the wild can't possible be 'server >>> systems'.
>> Huh? What? Where did you make that jump? I gave an example of why >> *one* system was clearly designed to be a server... not that all >> servers must be designed the same way!
>> At this point you are just trying to twist my words.
> I'm not. I'm just trying to figure out what you think is essential to > a 'server OS/system'. Throughout the discussion we have established > that it isn't the OS itself, but rather the server software and the > functionality it provides. That implies that I can turn e.g. XP into a > server OS/system. I gave the example of installing apache (any server > packages would've done really), but to you that was not enough. So I > an asking you what *is* enough. Since you can say what isn't, you must > be able to say what is too. Only, you can't seem to come up with an > answer.
I have responded to this and provided you with links and quotes to other responses. I have also talked about how there is no exact line to draw - noting there are gray areas.
>> Remember, the concept here is not complex:
> Indeed it isn't. When a system serves, it is a server. When it does > nothing but serving, it is a dedicated server. It has nothing to do > with the OS or the edition of the OS on the machine. The only thing > the 'server edition' of an OS does is give you a set of applications > that allow you to get started right away, and perhaps some > optimizations in e.g. the kernel and system settings.
Here you seem to understand that a server system has been not just been used as a server but also has been "optimized" for server use. Good.
>> In addition to the above being, to most experienced users, fairly >> common knowledge, I have pointed you to multiple resources to >> support this view. Your reaction is just to declare myself and the >> resources wrong and just repeat your claims that you are right with >> *no* support. None. Do you have any?
> I'm sorry, but you are lying again. I have given support for my claims > throughout the entire thread.
Your argument, which you restate below, is to say common terms are "meaningless". I have shown you where they have meaning. Other than for you to say they do not mean anything to *you*, even though they clearly mean something to others, what evidence do you think you have provided? What evidence could you? You would have to show how others are not understanding each other when they use the terms you claim are meaningless. If people *are* understanding each other then clearly the words have meaning. And they do.
The fact this gets you so worked up you would sink to making accusations of dishonesty does not speak well of you.
>> Not that it really matter... as I have noted, this is mostly just >> semantics. You are using uncommon terminology (something you have >> agreed you are doing) and just want to call others wrong for not >> using *your* terminology. You even go so far as to insist that >> other sources must be using your terminology in their descriptions. >> That is a bit silly of you.
> I did never insist on usage of any terms. I've noted that a lot of the > terms are meaningless and ambiguous, and they are.
They certainly are not black and white. But they have meaning... clearly.
>>> In addition to the above being, to most experienced users, fairly >>> common knowledge, I have pointed you to multiple resources to >>> support this view. Your reaction is just to declare myself and >>> the resources wrong and just repeat your claims that you are right >>> with *no* support. None. Do you have any?
>> I'm sorry, but you are lying again. I have given support for my >> claims throughout the entire thread.
> Your argument, which you restate below, is to say common terms are > "meaningless". I have shown you where they have meaning. Other > than for you to say they do not mean anything to *you*, even though > they clearly mean something to others, what evidence do you think > you have provided? What evidence could you? You would have to show > how others are not understanding each other when they use the terms > you claim are meaningless. If people *are* understanding each other > then clearly the words have meaning. And they do.
> The fact this gets you so worked up you would sink to making > accusations of dishonesty does not speak well of you.
It would not accuse you of dishonesty if you weren't dishonest. But you are. You say I make claims with no support. That's not true, hence it is a lie.
Here is me supporting my claims, over and over again:
"And I don't have to 'not accept that there is a difference'. There isn't a difference. Sure, Ubuntu's server edition ships with a different set of default packages and includes a menu in the installer where you can select different sets of server software packages, but in the end it's exactly the same operating system. I can turn an Ubuntu desktop installation into an Ubuntu server and vice versa in no time."
----
"The converting point I made was merely to enforce my opinion that there is no such thing as a 'server OS'. Any OS not marketed as a 'server OS' can just as much be a server (with any number of roles) as an OS that *is* marketed as a 'server OS'. There is no technical difference between the both."
----
"Sure, there are 'server' editions of particular operating systems, but that merely another way to say "if you buy this, you get a bunch of server software by default"."
----
"I know there are differences in OS distributions. A server edition tends to ship server software by default, while a desktop edition ships desktop software by default. That's the only difference. The OS is the same. And that's still exactly the point I was trying to make when I said that you can turn any OS in a 'server OS', which you wrongfully denied. It really doesn't matter *what* services are running; from the moment I declare a machine a server, it *is* a server, no matter what OS it happens to run."
----
"Why is it 'unwise' to run desktop software on a so called 'server OS'? I know some people running e.g. Windows Server 2003 as their desktop OS. I also know quite some people running a real wide open server (mostly game servers sporting http(s), SQL, IRC ... - some serving hundreds of users) on XP. They are doing just fine. You know why? Because you can turn any OS on a network into a server OS."
----
"I have never denied that some operating systems are released as a server OS. The only thing I'm saying is that the 'server' in 'server OS' describes the programs that come with the OS rather than the OS itself. Hence my opinion that there is no such thing as a server OS."
----
"But they are still the same OS. The 'desktop' and 'server' prefixes are just meaningless in regard to the OS itself, and just indicates what default set of packages is included."
----
"It's very odd by the way how you think the GUI to the postfix binaries is more determing in what makes a system a server than the postfix binaries themselves. It's common practice to run a server completely headless, and run graphical configuration interfaces (let's say WebMin) on a desktop client."
----
"You don't need to do any work at all to turn regular OSX into a 'server system'. It has apache. It has postfix. Being a Unix-like OS I pretty sure it has an FTP server and SSH by default as well.
None of the above services (of which any one is enough to set the machine up as a server) will perform better or offer more functionality on the server edition of OSX."
----
"'Web sharing'? Now what's that supposed to be? Another one of those terms that can mean anthing and nothing at the same time. Is this a web server? An FTP server? CIFS? NFS?
If you keep injecting meaningless terms, this discussion is going nowhere (not that it has been going anywhere at all of course).
There is no difference between serving on a lan and serving on a wan. It's just the scope that's different."
----
"Well, an FTP server facing the internet at large is 'web sharing' too."
----
"Well the common term for this kind of serving is 'web serving'. Why use yet another term?"
----
"I'd say I made a better server (system/OS/whatever) if I stripped off all but the bare essentials."
----
"Indeed it isn't. When a system serves, it is a server. When it does nothing but serving, it is a dedicated server. It has nothing to do with the OS or the edition of the OS on the machine. The only thing the 'server edition' of an OS does is give you a set of applications that allow you to get started right away, and perhaps some optimizations in e.g. the kernel and system settings."
|
