FreeBSDwiki - User contributions [en]

Xargs

2006-05-30T13:30:06Z

129.171.150.67:

from the manpage:
The xargs utility reads space, tab, newline and end-of-file delimited
strings from the standard input and executes utility with the strings as
arguments.

Any arguments specified on the command line are given to utility upon
each invocation, followed by some number of the arguments read from the
standard input of xargs. The utility is repeatedly executed until stan-
dard input is exhausted.

Talk:Apache, Installing with PHP

2006-03-29T16:21:40Z

129.171.150.67:

Are you a Linux guy, ColdFire? Just kinda curious as to why you'd compile Apache and PHP raw-dog rather than use the ports tree (and run a cron job to start it at boot time rather than use a /usr/local/etc/rc.d script). --[[User:Jimbo|Jimbo]] 01:49, 27 March 2006 (EST)

== not to mention ==

using [[wget]] instead of [[fetch]]
--[[User:Dave|Dave]] 13:00, 28 March 2006 (EST)

==Is There No Thing Called Free Speach?==
So is "from source" not welcome here? What is the point of me writing articles, only for you to just change the whole lot?

[[Installing_Apache_with_PHP]] was a "from source" guide, for one simple fact: You know where everything goes, and you can then easily query errors. And not like all, some people '''do''' prefer source installs -- which I'd thought I'd give the user a choice!

But it seems this a pkg_add based wiki. Great. What gives?

----
Actually around here we generally prefer the [[ports tree]], which Dave also gave the nod to. Which is also building from source, but has the (rather distinct) advantage of also creating entries in the ports database, as well as making sure that any FreeBSD-centric patches and file location changes get made, you can use portupgrade to update the port later, etc etc etc.

However I was still interested to see your "from tarballs" approach. It's not something I'd normally recommend on a FreeBSD machine when there is a port for that application, for the reasons mentioned above, but it's nice to see it done a bit. After all every now and then it is possible to run across something you want that there isn't already a port for!

Maybe you could write us an article specifically about building applications from tarballs rather than using the ports tree, and common gotchas and how to fix them?

PS: the talk page for your article is really a better place for this kinda thing, I had already asked you a leading question about some of this stuff there. =) --[[User:Jimbo|Jimbo]] 10:02, 29 March 2006 (EST)

User talk:Dave

2006-03-29T16:21:21Z

129.171.150.67:

==RE:welcome==
What do you think of the idea of putting the whole handbook in here and improving it through the wiki process? We may be able to get the doc project on board with that, especially since there are some tools (crude as of yet I think) that can convert mediawiki markup to docbook. Do you have the space and bandwidth for that? If it really takes off it could be merged back into the official FreeBSD webspace. - [[User:Taxman|Taxman]] 15:28, 17 Sep 2004 (GMT)

==problem with that==
There are two problems with that; a) this question is better asked of Jimbo (since this is his baby, not mine,) although I am pretty sure his answer's going to be along the lines of "no, I'd rather write something easier to deal with than to copy verbatim" and b) the handbook makes my eyes gloss over, and part of the reason I think a wiki is better than the handbook is because it brings fresher ways of presenting the information -- especially nice is the compactness of it (specifics when you want 'em). I mean, the handbook is already available in chunks of HTML-ized chapters....don't need to repeat work that's been done already. Although I'm sure Jimbo would be happy to see his work linked from/become part of official FreeBSD documentation (I know I'd be happy if someone saw an article I wrote and went "oh shit! that's how you do it! thanks!")

--[[User:Dave|Dave]]

==Dave said it right==
I'm not in favor of getting this project buried underneath the ''Handbook'', for all the reasons Dave listed. I get pretty frustrated every time I try digging information out of the ''Handbook''; so while I don't mind incorporating some of the stuff from it piecemeal, I am very very very much not in favor of dumping it here ''en masse''. --[[User:Jimbo|Jimbo]] 00:40, 18 Sep 2004 (GMT)

Running an SSH server in Windows

2006-03-28T17:58:35Z

129.171.150.67:

Why? So you can ssh into your Windows machines as well!

# Get Cygwin [[http://cygwin.com here]].
# Install it and choose the OpenSSH package (OpenSSL will also be chosen for you, as a dependency).
# Install the cygrunsrv package from the Admin category.
# Right-click "My Computer", select properties, go to Advanced tab, select "Environment Variables". Find the variable for "Path" and double-click it. Add "C:\cygwin\bin" to the path, after whatever else is there. Click on "New" and add a new variable named CYGWIN with the value set to ntsec tty.
# Open a cygwin bash shell and run ssh-host-config -y to set up ssh keys, your sshd config file and to set sshd to run as a Windows service. It'll ask for the value of the CYGWIN variable, put in the same thing as the var above: ntsec tty.
# start the service from the cygwin bash shell by running cygrunsrv -S sshd.
# You should be able to test it by ssh'ing to localhost. If that works, try it over your LAN. If that doesn't work, check your Windows firewall and allow connections to TCP port 22.

[[Category:Cygwin]]

Running an SSH server in Windows

2006-03-28T17:58:18Z

129.171.150.67:

Why? So you can ssh into your Windows machines as well!

# Get Cygwin <a href="http://cygwin.com">here</a>.
# Install it and choose the OpenSSH package (OpenSSL will also be chosen for you, as a dependency).
# Install the cygrunsrv package from the Admin category.
# Right-click "My Computer", select properties, go to Advanced tab, select "Environment Variables". Find the variable for "Path" and double-click it. Add "C:\cygwin\bin" to the path, after whatever else is there. Click on "New" and add a new variable named CYGWIN with the value set to ntsec tty.
# Open a cygwin bash shell and run ssh-host-config -y to set up ssh keys, your sshd config file and to set sshd to run as a Windows service. It'll ask for the value of the CYGWIN variable, put in the same thing as the var above: ntsec tty.
# start the service from the cygwin bash shell by running cygrunsrv -S sshd.
# You should be able to test it by ssh'ing to localhost. If that works, try it over your LAN. If that doesn't work, check your Windows firewall and allow connections to TCP port 22.

[[Category:Cygwin]]

FreeBSDwiki:Community Portal

2006-01-09T14:36:16Z

129.171.150.67:

What's this page for?

BSD-related links in general?

I wouldn't mind some BSD wallpapers.

----
Can someone here take the trouble and create /usr/ports/www/mediawiki?

----
I would appreciate some information in the commands section on process and file locking:
locating/tracking down locked files and processes. I would also appreciate some extra info
on what commands to use to unlock resources that remain locked after a program abnormally terminates.

-- Joe B. --

----
Can you give me a little more background info on this? I've never had any trouble with file or process locking; particularly given that, at least if you're root, there IS no such thing. Actually I'm unaware of per-user file locking on a non-superuser level either - which isn't to say that it doesn't EXIST, but I see an AWFUL lot of programs implementing their own flock system with special files to indicate locking on data files. Which tends to reinforce my (possibly naive) idea that if you want locking, you have to implement it for yourself..

So, yeah, can you give us more info about what you're having trouble with exactly?

--[[User:Jimbo|Jimbo]] 21:43, 17 Nov 2004 (EST)

----
Sounds like it's being init'd by root (or a user with root privs) at boot time and doesn't want to give it up. if you've used lsof to see what's open, kill -HUP or kill -9 it.

-dave

----
you can use [[lsof]] to see what files are open....

--[[User:Dave|Dave]]

----
I have installed Hylafax and tried to set it up. For some reason, the installation was interrupted and I keep getting the message that the modem [/dev/cuaa4], is 'locked', so no faxes can be sent. I can still dial the modem with a program like Seyon or Minicom. I've tried to hunt down the lockfiles used by the program, in case any of them are misbehaving, but with no luck and, besides, I'm still unsure how to unlock a locked socket or file if I find one. I'm running FreeBSD 5.1. The lockup continues even after a hard reboot and reinstall.

-- Joe B. --

----

Hrm. I'm unfamiliar with Hylafax, and to be honest my experience with modems under BSD is also nonexistent - I haven't used dialup under *nix since I got PPP set up under SuSE with my very first experiments with running my own *nix box about six years ago. What user context is your app running under? It's possible that you might have to do something special to delegate some privileges if you want it to run in a non-root context, sort of like mounting or dismounting CDs or floppies.

--[[User:Jimbo|Jimbo]] 16:08, 2 Dec 2004 (EST)

----

Hylafax has configuration files for its demons. They are similar to other allow/deny files, like, say, lpd's. All of those have been set up properly, as far as I can tell.

-- Joe B. --

check permissions on the daemons and the files they're supposed ot be accessing -- including devices. could be they don't have rights to do what they're supposed to and therefore fucking up.

Dave

----

Can we have a page for problem ports like [[java]] and eventually the walkthru for people who just left windoze, as well as a place to put a small page to present user?

Relax

----

not a bad idea -- i don't understand what you mean by &amp;quot;a small page to present user&amp;quot; since users of the wiki do have their own pages...click on jimbo's name (or [[User:Dave|Dave]]) for an example.

dave

----

Cool - but where is the page that points at the users?

Relax

----

[[Special:Listusers]]

btw, see the little icon at the top of the edit area that looks sorta like a signature, in between the red &amp;quot;nowiki&amp;quot; icon and the solid line icon? That's the signature icon, handy for Talk pages like this one, because it does this for you:

--[[User:Jimbo|Jimbo]] 13:03, 24 Dec 2004 (EST)

Should we post a story to slashdot so we can get more users?

[[User:Simon|relax]]

----

Fine by me. Hell, I'm advertising the thing on Google's adsense, what's wrong with ''free'' publicity? =) --[[User:Jimbo|Jimbo]] 01:19, 4 Jan 2005 (EST)

--------

My story of a FreeBSD wiki was rejected. I may be good with technical writing but not with marketing or story telling...

Someone else can give it a try? There are many slashdot editors so there is bound to be one that is thrilled by the idea of a BSD wiki!

We might want to provide one grandma-usable drolling-idiot-proof script that installs java on a browser or other feats unheard of in freeBSD newbie circles. It would prove technical competence, and provide a story because sun would stubornly insist on clicking on one of their agreements but made the whole install very obnoxious to most users precisely because they wanted users to see the license agreement. Such behavior isn't very open source minded.

Anyhow that's what I based my slashdot story attempt on (the obnoxious license preventing a simple &amp;quot;make install&amp;quot; from being used and the wiki as last hope of the FreeBSD newbie), but I'm sure someone else would word it better than me.

And THAT would be free publicity on a few sites, slashdot or no slashdot. (-;

P.S.: and we could all add a FreeBSD wiki link to every slashdot discussion about any form of wiki. Don't spam, just make sure you make a meaningful contribution to the discussion (i.e. be a karma whore and quote wikipedia first or make some kind of self-interview of your wiki experience. Pointing out which wiki codes are superior at what also helps).

[[User:Simon|relax]]

---------

I added a slashdot comment on the new java license fiasco story. It links here. Dear wiki, brace for impact!!

P.S.: we have spam in the main page at the bottom. How do wikis deal with those? Can we make the wiki auto-revert when certain spamlinks are added??

[[User:Simon|relax]]

--------

I saw you link to the wiki, and I added you as a &amp;quot;friend&amp;quot; (jimbo's on my friends list as well,) and my story to OSNews.com didn't get published, so it's fair to say that it was rejected as well.

The main problem with a simple &amp;quot;one script to install Java on FreeBSD&amp;quot; is that a) such a script would break the Sun License w/r/t Java -- this is the main reason there isn't a port for it, AFAIK -- and b) such a script would have to be really fucking complex to take into account all the different FreeBSD distros (4 series, 5.x series, various architectures etc) as well as checking that all the prereq's are available on your system. Also, there's no simple way to install all the files you have to download since you a) have to have a login to get to the Sun download page and b) have to get a bunch of stuff. If you were just able to download the files, a simple fetch or wget would suffice, but as it is, you have to login, agree to a license agreement and select which downloads to get. Downloading the files in to one central location is not an option for the sun license agreement(s). Sun might not be as big as Microsoft or as popular as Apple, but they're a big company that would be Very Unhappy if someone were to distribute their stuff without their consent. Large companies with lots of lawyers and piles of money are Not Fun when you piss them off.

The closest you could really get is a script that printed &amp;quot;you need to get file12312-sadf-ix86.tar.bz2 before you can continue. Do you want to do this? You will need to login and download the file manually.&amp;quot; and if the user enters &amp;quot;yes&amp;quot;, do a &amp;quot;mozilla http://downloads.sun.com/java&amp;quot; or something. Added to the fact that it's not a sure thing that a given computer will have a GUI, much less a browser, much less mozilla specifically. If it were easy, the FreeBSD.org folks who do Java (and there seem to be a fair number of them,) would have done so. I'm not saying it's impossible, but it's a big PITA. You may be able to compile a package and distribute THAT, but I doubt it; I'm fairly sure that the blackdown and other java package maintainers are in the FreeBSD committers group(s) and as such are covered by the FreeBSD/Sun Java agreement.

--[[User:Dave|Dave]] 21:03, 7 Jan 2005 (EST)

---

I never mean a port that illegally installs java without displaying a licence!!

I mean a script - or some scripts - to do the technical grunt work those ports are leaving clueless users to do and point you exactly your errors as you make them so to avoid the 160+ hours of non-license-related technicalities I've been going thru as a simple user.

The java install process is very far from smooth. I've learned tons of FreeBSD technicalities along the way, but I'm no closer to a working java in any browser than I was last week or the week before.

I'd gladly install WINE and internet explorer on top if it would get java to work, but apparently this port is having config difficulties that annoys the pros so I won't try that...... yet.

And sure, the license part is annoying and manual fetching of files at URL that no longer exist (version number changes - 10 minutes of aggravating guesswork) repeated 6 times as the port won't list all your needs at once is annoying but it's nothing compared to the actual install or linking to browser which is what we should have a port for.

A &amp;quot;how to intall it&amp;quot; documentation port would be welcome if even a script pointing you the newbie-invisible mistakes that don't leave error messages would have legal problems (I don't see how).

As for the varieties of BSDs out there and browsers, why not have a port that does just any browser the port-maker prefers and the latest BSD as an example of what to do? I haven't seen a complete walkthru for any browser whatsoever (so I'm never sure if the 7 incomplete install instructions I have all get the same omission). The recent license pull by JAVA has even affected my option to go back for old binaries...

And sorry for being french, what's &amp;quot;PITA&amp;quot;?

[[User:Simon|relax]]

----

PITA = Pain In The Ass. And installing Java is ALWAYS going to be a PITA as long as Sun requires you to click a very complex piece of shit in a web page that can't possibly be rendered without a complex graphical browser in order to download the code, which is one reason why I very very frequently rather fervently espouse AVOIDING Java - the damn thing just isn't that great of a platform to make it worth saddling yourself with all that legal shit, &amp;quot;free&amp;quot; or not. It might be free as in beer, but it's a loooooooooooong way from being free as in speech.

As far as putting together a *complete* walkthrough... well, that's what the wiki here is all about. By all means, hit it up! You're right, a complete walkthrough of getting Java set up for [browser] would be great, and helpful... so make one. That way, the NEXT time you have to do this, you'll have a nicely documented path to walk down of all the things that actually worked the first time - which is, again, what this wiki is all about. =) If all the good simple documentation were already out there, I wouldn't have set this thing up to begin with. --[[User:Jimbo|Jimbo]] 22:59, 12 Jan 2005 (EST)

---

I added this wiki to my stumpleupon page. Those moz extensions (stubleupon being my fav) are truly something M$ can't duplicate! (-;

[[User:24.202.242.123|24.202.242.123]]

----

Domain has no postmaster!! I am getting this in vqadmin Any idea

(62.254.0.55)

----

Well, 62.254.0.55, you aren't giving us much to work with. Is this a brand new server with no previously configured domains, or were you already serving mail for the domain in question on the server in question, and now you're trying to install vqadmin on the top of it? --[[User:Jimbo|Jimbo]] 15:14, 13 Apr 2005 (EDT)

----

Talk:Ports, Installing

2005-11-21T14:43:25Z

129.171.150.67: yeah, it should be more cohesive

==just btw...==
you have duplicated effort:

http://freebsdwiki.net/index.php/Searching_ports

http://freebsdwiki.net/index.php/Update_the_ports_tree

--[[User:Dave|Dave]] 22:50, 15 Jan 2005 (EST)

Shouldn't there be some mention of:
1) CVS
2) Building the index (since 5.3 index file has been removed).
3) Installing ports

== yeah, it should be more cohesive ==

there are entries for [[cvsup]], [[portupgrade]] and some troubleshooting ports stuff, but yeah...really should clean up the ports section so that it's not like 5 different articles. --[[User:129.171.150.67|129.171.150.67]] 09:43, 21 Nov 2005 (EST)

Tcpdump

2005-03-24T18:41:14Z

129.171.150.67: /* Common arguments */

'''tcpdump''' is an extremely handy little utility that will "sniff" traffic on a particular network interface, match it against a set of criteria, and then output a summary of it to screen (or a dump of it to a file).

==Common arguments==

-c ''number'' "sniff this many packets". By default, tcpdump will sniff until
you tell it not to. Using this flag will cause it to stop at a certain number
of packets instead of waiting for an interruption (like cntrl-C...or running
out of diskspace, if you're writing to a file).
-i ''interface''"listen on this interface". If you want traffic from just one
network interface, you want to use this option.
-a "convert IPs to names if you can"
-w ''filename'' "write this to a file". you must give a filename to write to.
-q "don't be so verbose" -- strips more protocol information
-p "do not put interface in promiscuous mode". By default, tcpdump will make your
interface promiscuous.
Note that if not src (source) or dst (destination) port/ip/net is specified, tcpdump assumes ''either'' source or destination:
src port ''number'' "from this port"
dst port ''number'' "to this port"
src ip ''number'' "from this IP"
dst ip ''number'' "to this IP"
src net ''number'' "from this network"
dst net ''number'' "to this network"
[[tcpdump]] understands boolean operators (and not or, etc) and can take hostnames, IPs, networks and protocols as arguments. The output is terse and hard to understand if you don't know what you're looking at or for; for this reason many folks prefer friendlier front-ends to tcp, such as ethereal.

==ok, but is it useful?==

Just tonight, I discovered that while I had login information for an FTP site I hadn't touched in years saved in my Windows FTP client, I didn't remember what that password WAS, and I needed it to give to somebody else. I looked in the data store for the FTP client, but unfortunately the passwords were stored as hashes. Rather than try to figure out the encryption algorithm used and then find a brute force cracker somewhere, I just shelled into my [[firewall]] and set up a '''tcpdump''' session to capture the packets:

ph34r# '''tcpdump -pw /home/jimbo/ftpsniff.tcpdump.bin -i xl0 dst port 21'''
tcpdump: listening on xl0

Now my firewall is monitoring all traffic through its inside interface, and dumping any packets headed out to an FTP server to the file '''/home/jimbo/ftpsniff.tcpdump.bin'''. So I fire up my Windows FTP client, connect to the FTP site and let it authenticate, and then immediately hit CTRL-C in ph34r to interrupt the tcpdump session:

ph34r# '''tcpdump -pw /home/jimbo/ftpsniff.tcpdump.bin -i xl0 dst port 21'''
tcpdump: listening on xl0
^C
38 packets received by filter
0 packets dropped by kernel

And hex editing the raw tcp dump (I cheated and used UltraEdit on my windows box; I don't have a good hex editor for *nix yet. Anyone got any recommendations?) and searching for PASS led me to this snippet:

PASS V9xo3Pr1

Ta-da! Now I know the password for that FTP site. (Obviously you could use the same technique to find out OTHER people's passwords, if you ran the firewall they used to get out to the 'net... starting to see why the crypto geeks keep yelling about using secure authentication instead of plaintext?)

[[Category:System Commands]]

Bash

2005-02-10T19:25:27Z

129.171.150.67:

The Bourne Again Shell (located in /bin/bash) is the default shell of the [[Linux]] operating system and is the shell that users of that system will likely be most familiar with.

One of bash's strongest features, shared with the [[Bourne shell]], is flexible output [[redirection]].

Note: Bash is ''not'' available by default in the base system, but can easily be installed from [[:Category:Ports and Packages|ports]] if desired.

See [[bash]]'s [[man]] page for more info.

see also: [http://cnswww.cns.cwru.edu/php/chet/bash/bashtop.html bash homepage]

To change your shell from one to another, run the [[chsh]] command.

To change [[bash]]'s look and feel, edit your [[shell configuration file]] -- .profile and/or [[.bashrc]] (may be called .bash_profile on older systems).

Other shells that you can install and customize for ease of use are the [[bash]], [[tcsh]], [[psh]], [[ksh]], [[zsh]].

See also [[Changing_your_shell]]

[[Category: Shells]]
[[Category: Ports and Packages]]

USB storage

2004-12-27T19:19:16Z

129.171.150.67:

Okay, I just mounted my own 512MB Lexar JumpDrive (USB keychain storage device) on ph34r (my FreeBSD 5.2.1 amd64 box) for the first time. It was pretty easy!

First, I plugged in the JumpDrive. Since I was shelled in instead of logged in at the console, I couldn't see the console messages that pop up when you change devices. So the next thing I did was check the end of /var/log/messages for the kernel messages about the newly attached JumpDrive:

ph34r# '''tail /var/log/messages'''
Sep 9 14:18:09 ph34r kernel: umass1: LEXAR MEDIA JUMPDRIVE, rev 2.00/20.00, addr 3
Sep 9 14:18:09 ph34r kernel: GEOM: create disk da4 dp=0xffffff003c46e068
Sep 9 14:18:09 ph34r kernel: da4 at umass-sim1 bus 1 target 0 lun 0
Sep 9 14:18:09 ph34r kernel: da4: <LEXAR JUMPDRIVE 2000> Removable Direct Access SCSI-0 device
Sep 9 14:18:09 ph34r kernel: da4: 1.000MB/s transfers
Sep 9 14:18:09 ph34r kernel: da4: 493MB (1010784 512 byte sectors: 64H 32S/T 493C)

Great! It auto-discovered fine, and it was detected as device '''da4'''. Knowing that, let's check to see what we've got in the way of disk [[partitions]] and [[slices]] on the drive:

ph34r# '''ls /dev | grep da4'''
da4
da4s1

Okay, cool - it's a simplified system, one partition, unsliced. No problem. As standard for USB mass storage devices, I already know this device is going to have a FAT file system on it - but if I wasn't sure, I could of course have plugged it into a Win2K/XP machine, gone to Disk Manager, and checked the properties on it. You never know, ONE day they might start making these things with NTFS as the standard filesystem... but I wouldn't hold my breath, if I were you. Anyway, yes, it's FAT, so off we go with making a directory to mount it on and mounting it:

ph34r# '''mkdir /mnt/usbdrive'''
ph34r# '''mount -t msdos /dev/da4s1 /mnt/usbdrive'''
ph34r# '''ls /mnt/usbdrive'''
FireFox 0.9.2 + MPClassic Partition Magic 8.0
Malware Removal Undelete
NFS-Server(true-grid-pro)1.0 NT - 2K password reset util
putty.exe OpenOffice 1.1.2
OpenVPN-Win32 vnc-4.0-x86_win32.exe
vnc-4.0-x86_win32_viewer.exe
ph34r#

Tada! Everything's mounted, and I can see all of my work-related files are in there just where they're supposed to be. Now I'll just add an entry to /etc/fstab to make mounting it a little easier next time:

ph34r# '''cat /etc/fstab'''
# Device Mountpoint FStype Options Dump Pass#
/dev/ad4s1b none swap sw 0 0
/dev/ad4s1a / ufs rw 1 1
/dev/ad4s1e /tmp ufs rw 2 2
/dev/ad4s1f /usr ufs rw 2 2
/dev/ad4s1d /var ufs rw 2 2
/dev/ad6s1e /data ufs rw 2 2
/dev/acd0 /cdrom cd9660 ro,noauto 0 0
/dev/da4s1 /mnt/usbdrive msdos rw,noauto 0 0

And voila! Now from here on out I can just '''mount /mnt/usbdrive''' and '''umount /mnt/usbdrive''' without issues. Notice that for the USB drive, I set FStype to msdos, Options to rw (read and write) and noauto (do not automatically attempt to mount the volume at boot time - the media won't always be present!), and set dump and pass to 0 each, to tell the system also not to try to automatically [[fsck]] the USB drive at boot time either.

One final note: don't forget to [[umount]] the USB drive before you disconnect it - you could lose data or even leave the filesystem itself with some problems if you just yank the drive out and it turns out that there were still cached writes pending that you didn't know about.

[[Category:Common Tasks]]

[[Category:FreeBSD for Workstations]]