Date: Fri, 7 Aug 2009 13:08:19 -0700 (PDT) From: James Phillips <anti_spam256@yahoo.ca> To: freebsd-questions@freebsd.org Subject: Re: FreeBSD for the common man(or woman) (was: > upgrade 7.2 Message-ID: <457708.23496.qm@web65507.mail.ac4.yahoo.com>
next in thread | raw e-mail | index | archive | help
Thank-you for your replies. I guess=0Amy main concern was I'm not sure when= to stop banging my=0Ahead against the wall and ask for help. The checklist= kind=0Aof goes like:=0A=0ADid you read the FAQ and release notes?=0ADid yo= u read the handbook?=0ADid read the man pages?=0ADid you search the mailing= -list archives?=0A=0AThis list is probably best suited to very specific=0Aq= uestions.=0A=0ASome the stuff I mentioned has little to do with BSD.=0A=0A-= -- On Thu, 8/6/09, Polytropon <freebsd@edvax.de>=0Awrote:=0A=0A> From: Poly= tropon <freebsd@edvax.de>=0A> Subject: Re: FreeBSD for the common man(or wo= man)=0A(was: > upgrade 7.2=0A> To: "James Phillips" <anti_spam256@yahoo.ca>= =0A> Cc: freebsd-questions@freebsd.org=0A> Received: Thursday, August 6, 20= 09, 7:45 PM=0A> On Thu, 6 Aug 2009 14:56:41 -0700=0A> (PDT), James Phillips= <anti_spam256@yahoo.ca>=0A> wrote:=0A<SNIP!>=0A> =0A> =0A> =0A> =0A> > I c= hecked the HP website: they will release the=0A> details of the PCL=0A> > l= anguage (version 4 or so) for a price.=0A> =0A> The PCL language is usually= output by gs (the Ghostscript=0A> printer=0A> "driver" collection that tra= nslates PS into PCL and other=0A> printer=0A> languages).=0A> =0AYes, I fig= ured this out when I abandoned the Handbook and=0Alooked at the ports colle= ction. I've used GhostScript under=0Awindows as well.=0A> =0A> =0A> > I fin= ally got it working by installing the Apsfiler=0A> package in the=0A> > por= ts collection (no, did not send the post-card yet;=0A> the print=0A> > serv= er is not functional yet.)=0A> =0A> Personally, I prefer apsfilter to CUPS,= but maybe you would=0A> have liked=0A> CUPS better. It offers a browser ba= sed interface and offers=0A> lots of=0A> autodetection functionality. (But = you can't install a=0A> parallel printer=0A> that isn't connected to the sy= stem easily, for example.)=0A> =0AI'm wondering how well apsfiler and CUPS = cooperate. Samba=0Auses CUPS by default.=0A=0A> Setting up a printer with t= he apsfilter SETUP script is=0A> very easy as=0A> long as you know which na= me the printer has - you mentioned=0A> HP. And=0A> if it's a HP Laserjet, y= ou're lucky. You're even more lucky=0A> if your=0A> printer does support th= e PS standard, because then you can=0A> avoid using=0A> any printer filter = (such as apsfilter) because PS is the=0A> default output=0A> format for pri= nting, and it can be fed directly into the=0A> printer.=0A> =0A=0AThe print= er is a $10 POS I got used. PS output seems to=0Aconfuse it. I'm tempted ju= st to get a newer one.=0A=0A(Laserjet 5L -> except it gets confused by PCL = 5 as well)=0A> =0A> =0A<SNIP!>=0A=0A>=0A> > made few tweaks of the system t= o better follow the=0A> Filesystem=0A> > Hierarchy Standard (such as symlin= king /usr/local/etc=0A> to /etc/opt).=0A> =0A> Erm, excuse me? First of all= , it's not encouraged to mix OS=0A> things=0A> with application things. You= know that FreeBSD keeps the=0A> difference=0A> between "the OS" and "every= thing else" (which is=0Alocated in=0A> the=0A> /usr/local subtree). If you'= re coming from a Linux=0A> background, I=0A> could understand that you're n= ot familiar with this=0A> concept.=0A> The /usr/local subtree can be comple= tely removed and=0Astill=0A> leaves=0A> you with a completely intact and fu= nctional OS.=0AEverything=0A> that=0A> you install by ports or packages goe= s into /usr/local,=0Aand=0A> of course,=0A> the configuration files belong = there, too.=0A/usr/local/etc=0A> has the=0A> same structure as /etc, but it= 's reserved for=0Aadditional=0A> software.=0A> Vice versa, configuration fi= les of locally installed ports=0A> do not=0A> belong into /etc.=0A> =0A> Re= fer to =0A> =0A> =A0=A0=A0 % man hier=0A> =0A> to learn where things are ke= pt on FreeBSD.=0A> =0AUsing a symlink (/usr/local/etc -> /etc/opt) , the=0A= system IS still functional if /local is not mounted.=0A=0Aputting the setti= ngs in /etc makes it possible to mount=0A/usr read-only (in theory).=0Ahttp= ://www.pathname.com/fhs/2.2/=0AAccording to Wikipedia, it is Linux-specific= .=0A=0AIn any case, the changes are minor.=0A> =0A<SNIP!>=0A> =0A> > That m= achine is slowly degrading while I try to get=0A> the fileserver=0A> > work= ing the way I want.=0A> =0A> That indicates a major problem. Either your ha= rdware is=0A> faulty, or you=0A> are treating the software in the wrong way= .=0A=0A"Machine" in that sentence refers to win98 client. The HD=0Aactivity= light stays on for no apparent reason (no=0Athrashing). I suspect malware,= even if the Anti-virus can't=0Afind it :(=0A> =0A> =0A> =0A> > The last ti= me I did a complete re-install (of win98) I=0A> lost data=0A> > due to a da= maged disk that I copied the data to (and=0A> learned that=0A> > bzip2recov= er is a quick hack that needs to be=0A> re-written properly=0A> > according= to the source code).=0A> =0A> It's completely normal that you lose data on= "Windows"=0A> platforms.=0A> That's why you have a UNIX server for backups= .=0A=0AI lost data under Linux that I used for doing the back-up.=0AI blame= the cryptic HD error messages under Linux. Took=0Ayears to figure out what= happened. I think FreeBSD=0Ahard-drive failure messages were less cryptic.= =0A> =0A> =0A> =0A> > I hope to replace windows with wine for the most part= ,=0A> but wine=0A> > simply installs the applications in the users' home=0A= > directory=0A> > (breaking the FHS).=0A> =0A> No. You run wine as a user a= pplication, so you have=0A> user=0A> rights only.=0A> Then, wine of course = provides a user-based installation of=0A> your=0A> desired "Windows" progra= m.=0A=0AAn analogy would be to think of Wine like a generic=0Ainterpreter l= ike the BSD Linux compatibility, or the Java=0Aruntime for running Java app= lications. There should be a way=0Afor the administrator to install windows= applications in=0Asomething like /usr/local.=0A=0AThis would only be on th= e "client" machines anyway (not the=0Aserver). I Just hear a sucking sound = every time I try to=0Ainstall a 500MB+ game in the user directory.=0A=0A> = =0A> A workaround could be to create a special user, e. g.=0A> "wincompat",= =0A> that is used to install the applications. His home=0A> directory=0A> /= home/wincompat is then set to allow access for other=0A> users, so=0A> they= can execute the (there installed) "Windows" programs=0A> instead of=0A> re= quiring a user-local installation.=0A> =0A> =0A> =0A> > This is only resolv= able IMHO by having wine use a real=0A> database=0A> > back-end for the reg= istry (allowing user-level "views"=0A> of the=0A> > data, while still isola= ting different users).=0A> =0A> That's a bit complicated, but still possibl= e with the=0A> standard=0A> file attributes on a UFS file system - I hope I= understood=0A> you=0A> correctly.=0A=0AWine uses a flat file (for the regi= stry). I don't think it=0Asupports per-user isolation, so each user has the= ir own wine=0Aenvironment instead. Again Wine, not BSD related.=0A> =0A> = =0A> =0A> > Setting up NFS was a lesson in the intecracies of NIS=0A> twice= =0A> > since my Linux clients do things a little differently.=0A> After=0A>= > asking on one of the IRC channels that we are not=0A> advised to use; [.= ..]=0A> =0A> We?=0AGeneral public. Most the "supported" IRC channels requir= e=0Aauthentication.=0A> =0A> =0A> =0A> > [...] I edited the /var/yp/Makefil= e to suppress groups=0A> outside=0A> > the range of (1001 -2000). That basi= cly prevents the=0A> "special"=0A> > groups from being exported to the Linu= x clients (that=0A> use different=0A> > numbering)=0A> =0A> It's traditiona= l to create a "name:name" for user "name".=0A> Of course,=0A> there's no pr= oblem of grouping your user names to=0A> "name:group1" and=0A> similar thin= gs; "name:staff" comes into mind.=0A=0Ayes, the user groups are in the 1001= + range. this is=0Aconsistent with what Debian does. The "special" groups = like=0Appp, audio, etc, are numbered differently. NFS only passes=0Auser an= d group numbers over the network. So, to keep things=0Aremotely sane, NIS i= n needed to keep all the user names=0Aconsistent. I originally didn't want = to use NIS.=0A=0A> =0A> =0A> =0A> > To do this, I DID need the gory low-lev= el details in=0A> the handbook.=0A> > I didn't note the exact date, but I r= eally didn't=0A> touch the server=0A> > for months after that. I copied my = work to the Linux=0A> client because=0A> > the hard-disk was failing, and I= still did not get=0A> DVD-burning=0A> > working.=0A> =0A> There's the grow= isofs tool from the ports. The port's name=0A> is=0A> dvd+rw-tools. You sim= ply run=0A> =0A> =A0=A0=A0 % growisofs -Z /dev/dvd -r -J somedata/=0A> =0A>= or refer to=0A> =0A> =A0=A0=A0 % man growisofs=0A> =0A> for detailed infor= mation. (I have a symling dvd -> cd0=0A> so it fits to=0A> the manpage.) Of= course, you don't need ISO-9660 for data=0A> DVDs, you=0A> could record ta= r archives onto them directly, but that's=0A> another topic.=0A> =0A=0AI'll= look into that. I was planning on using dump/restore=0Afor most the system= , then using tar for selected directories=0Ain /var.=0A=0A> =0A> =0A> > At = one point when doing a Google search for "fxp" I=0A> came across this messa= ge:=0A> > http://unix.derkeiler.com/Mailing-Lists/FreeBSD/current/2008-10/m= sg00340.html=0A> > Call for testers: fxp(4) WOL=A0 <- My card!=0A> =0A> The= fxp NIC is an Intel one, right? I still have one, works=0A> good, and=0A> = is excellently supported.=0A=0Ayes. Mine is old though (Intel 82558)=0A> = =0A> =0A> =0A> > At that point, I decided to install the FreeBSD=0A> testin= g release=0A> > (7.x).=0A> =0A> Testing? A release isn't for testing, it's = for a productive=0A> and functional=0A> system. If you intend testing, you = would require the latest=0A> sources from=0A> HEAD, which is 8-CURRENT at t= he moment.=0A=0AIt went stable by the time I installed it.=0A> =0A> =0A<SNI= P!>=0A> =0A> > I finally installed FreeBSD 7.2 (release) on May 9,=0A> 2009= . However, I=0A> > now note some feature creep:=0A> > In addition to file/p= rint and backup server, I want=0A> to:=0A> > 1. Have it sleep when not in u= se (part of the delay=0A> was figuring out=0A> > how to get the router to s= end the magic packet. I read=0A> RFC's to=0A> > determine the proper way, a= nd found a "hack" that will=0A> work on my=0A> > floppy-based router for my= network set-up (send it=0A> every DHCP lease).=0A> Are you talking about t= he "wake on LAN" feature?=0AYes.=0AMy old NIC requires a specially-crafted = packet to wake up.=0A> =0A>=A0 =0A> > 2. I think I want to move the Voice/F= ax/Modem to the=0A> machine. Recently=0A> > I realised a lack of WakeOnRing= may impair phone answering if machine=0A> > is sleeping.=0A> =0A> I think = that's been calling "wake on modem"... it reminds=0A> me to a funny=0A> acc= ident when a customer told me that when his phone rings,=0A> his computer= =0A> starts up. :-)=0A=0AI'm hoping the serial port interrupt will be enoug= h.=0A> =0A> =0A> =0A> > 3. I still hope to do "other things" once the machi= ne is working=0A> > reliably.=0A> FreeBSD's good load management should mak= e this possible.=0A> =0AMay need to compile a custom kernel to free some me= mory.=0Aavail memory with generic: 237 MB of 256.=0A=0ANot a priority thoug= h.=0A> =0A> =0A<SNIP!>=0A> > I hope with the server properly set up, win98 = may even be usable=0A> > again: just do a clean install every morning!=0A >= Doesn't it reinstall automatically by itself? :-)=0A=0AOver a 100Mbps netw= ork, a 700MB image will take over a=0Aminute to transfer.=0AActually, that = is not bad.=0A =0A<Snip!>=0A=0ARegards,=0A=0AJames Phillips=0A=0APS: forgot= to cc the list.=0A=0A=0A=0A _________________________________________= _________________________=0AYahoo! Canada Toolbar: Search from anywhere on = the web, and bookmark your favourite sites. Download it now=0Ahttp://ca.too= lbar.yahoo.com.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?457708.23496.qm>