Re: ezjail
Thanks Seph, it works! Regards Matthias On 11.05.17 05:23, Sepherosa Ziehau wrote: > On Thu, May 11, 2017 at 12:42 AM, Matthias <matthias_p...@gmx.net> wrote: >> Hi, >> >> found that the crash was introduced with commit related to >> multi-threaded UDP traffic handling. >> >> When reverting changes done in that commit in latest CURRENT the crash >> does not occur anymore and ezjail works without problems. >> >> Filed a bug report: >> https://bugs.dragonflybsd.org/issues/3035 > > It should have been fixed by: > 8280d8f175b06023d940cba972a2c0b714fa21e3 > > Thanks, > sephe >
Re: ezjail
Hi John, thank you for the quick answer. I installed ezjail over two years ago and updated all packages regularly (with pkg upgrade). This time I deleted all packages to reinstall them since I had issues with the server. ezjail had been successfully running in DragonFly CURRENT from end of March this year. After upgrading two weeks ago I got a core dump during jail starts, so it seems that is due to your mentioned missing libjail.so. Is there an alternative that is also as equally easy to setup? Regards Matthias On 07.05.17 14:34, John Marino wrote: > "anymore" has been a year. > IIRC ezjail requires libjail.so which is specific to FreeBSD -- In other > words ezjail built but didn't work. > > Are you sure it's functional on DragonFly? > > John > > On 5/7/2017 06:47, Matthias wrote: >> Hi, >> >> I wanted to upgrade my server recently and discovered there is no >> package for ezjail anymore which I use for several jails running on this >> server. >> >> Will the package be available again anytime soon or should I install it >> from ports? >> >> >> Regards >> Matthias >>
ezjail
Hi, I wanted to upgrade my server recently and discovered there is no package for ezjail anymore which I use for several jails running on this server. Will the package be available again anytime soon or should I install it from ports? Regards Matthias
Re: Running firefox a bit more safely - HOWTO
Ok, changed :0 to $DISPLAY and it works now. Thank you! Regards Matthias On 26.09.16 20:22, Matthew Dillon wrote: > Ah, then that script won't work as-is. you'll have to record what your > current $DISPLAY is somewhere (do it in the same script that scp's your > .Xauthority) and then setenv DISPLAY to that value in the ssh that runs > firefox). > > -Matt
Re: Running firefox a bit more safely - HOWTO
Thanks for your quick response. I noticed that my default display is not fixed at :0, but I have :45 at the moment, need to resolve that. Regards Matthias On 26.09.16 19:33, Matthew Dillon wrote: > Well, running through the ssh tunnel itself is going to be really slow. > It will work, but it won't be fun. Make sure your *current* .Xauthority > file is installed in the other account and also make sure the other > account is in the 'video' group. .Xauthority changes every time you > start X. I run a little script to install it whenever I start X (after > a fresh reboot of my workstation, for example). > > -Matt > > On Mon, Sep 26, 2016 at 9:41 AM, Matthias Play <matthias_p...@gmx.net > <mailto:matthias_p...@gmx.net>> wrote: > > Hi Matt, > > the script does not work for me as it is listed. I need to call ssh with > '-Y' to get it working. > > I get the following error when calling ssh like it has been suggested: > > "Unable to init server: Could not connect: Abstract UNIX domain socket > addresses not supported on this system > Error: cannot open display: :0.0" > > What might be the reason for this? > > Regards > Matthias > > > On 12.08.15 20:32, Matthew Dillon wrote: > > The sshd sets the DISPLAY environment variable to point to its tunnel. > > There is no requirement that you use the tunnel, hence the above > script > > overrides it and sets the DISPLAY to :0.0 (which is a direct local > > connection) before running firefox. > > > > -Matt > > > > On Wed, Aug 12, 2015 at 2:22 AM, Carsten Mattner > > <carstenmatt...@gmail.com <mailto:carstenmatt...@gmail.com> > <mailto:carstenmatt...@gmail.com <mailto:carstenmatt...@gmail.com>>> > wrote: > > > > On Tue, Aug 11, 2015 at 7:32 PM, Matthew Dillon > > <dil...@apollo.backplane.com > <mailto:dil...@apollo.backplane.com> > <mailto:dil...@apollo.backplane.com > <mailto:dil...@apollo.backplane.com>>> > > wrote: > > > #!/bin/csh > > > # > > > # script for ~/bin/firefox (assumes ~/bin is in > your path) > > > # > > > scp ~/.Xauthority dfw1@localhost: > > > ssh dfw1@localhost -n "setenv DISPLAY :0.0; firefox" > > > > snip > > > > > * Also note that these applications will be able to use > X shared memory > > > and thus run fairly optimally (they are NOT using an > ssh tunnel nor do > > > we want them to as that would be ridiculously slow). > Programs run in > > > this way will not have direct access to the GPU so 3D > might not be > > > so hot. But for general browsing I haven't had any > trouble, and even > > > with 4K video appears to work about the same as it did > before. > > > > Can you explain how this works without going through localhost:22? > > Isn't X forwarded through the ssh tunnel? > > > > > >
Re: Running firefox a bit more safely - HOWTO
Hi Matt, the script does not work for me as it is listed. I need to call ssh with '-Y' to get it working. I get the following error when calling ssh like it has been suggested: "Unable to init server: Could not connect: Abstract UNIX domain socket addresses not supported on this system Error: cannot open display: :0.0" What might be the reason for this? Regards Matthias On 12.08.15 20:32, Matthew Dillon wrote: > The sshd sets the DISPLAY environment variable to point to its tunnel. > There is no requirement that you use the tunnel, hence the above script > overrides it and sets the DISPLAY to :0.0 (which is a direct local > connection) before running firefox. > > -Matt > > On Wed, Aug 12, 2015 at 2:22 AM, Carsten Mattner > <carstenmatt...@gmail.com <mailto:carstenmatt...@gmail.com>> wrote: > > On Tue, Aug 11, 2015 at 7:32 PM, Matthew Dillon > <dil...@apollo.backplane.com <mailto:dil...@apollo.backplane.com>> > wrote: > > #!/bin/csh > > # > > # script for ~/bin/firefox (assumes ~/bin is in your path) > > # > > scp ~/.Xauthority dfw1@localhost: > > ssh dfw1@localhost -n "setenv DISPLAY :0.0; firefox" > > snip > > > * Also note that these applications will be able to use X shared > memory > > and thus run fairly optimally (they are NOT using an ssh tunnel > nor do > > we want them to as that would be ridiculously slow). Programs > run in > > this way will not have direct access to the GPU so 3D might not be > > so hot. But for general browsing I haven't had any trouble, and > even > > with 4K video appears to work about the same as it did before. > > Can you explain how this works without going through localhost:22? > Isn't X forwarded through the ssh tunnel? > >
ipfw3 show dumps core after adding forwarding rule
Hi, after adding the following rule to ipfw3 rule set 'ipfw3 show' dumps core: ipfw3 add forward tcp 192.168.128.5:22 to 192.168.128.1:22 Regards Matthias
IPFW3 problems with network lists
Hi,
I use a shell script to setup my ipfw3 firewall and want to define
network lists to shorten my rule set. For that I used the approach
described inside the example section on ipfw3 in 'man ipfw3' like the
following:
#!/bin/sh
nets_allowed="{ 192.168.1.1 or 192.168.1.2 }"
ipfw3 add allow log 1 ip in from ${nets_allowed} to 192.168.0.1
When I run this script I get the following error:
ipfw3: hostname ``{'' unknown
Can you also reproduce this error?
Regards
Matthias
Re: IPFW3 and IPv6
Hi Bill, thanks for your quick answer! An NFS daemon is running on my server machine which I use to mount DFly source and obj dirs on my client machines. With NFS comes mountd and rpc* daemons that automatically also provide their service on IPv6 ports. I did not find daemon flags to tell them only to use IPv4 so I would like to block unwanted traffic to these ports. Would it be ok to also run ip6fw alongside with ipfw3? If yes, I will set up a configuration file for ip6fw allowing all IPv4 traffic and blocking all IPv6 traffic. Regards Matthias On 03.01.16 14:09, bycn82 wrote: Hi Happy New Year Matthias, not support IPv6 yet but yes i want to integrate with ip6fw as well. recently working on something related to voip. so did not find much time for BSD. Will support v6 once I remove the lock in NAT. Regards, Bill Yuan On Sunday, 3 January 2016, Matthias Play <matthias_p...@gmx.net <mailto:matthias_p...@gmx.net>> wrote: Hi, Happy New Year and all the Best in 2016! I set up ipfw3 to filter network traffic on a local server and want to filter all IPv6 traffic since I do not need it in my local network. Is ipfw3 also filtering IPv6 traffic or only IPv4? I am wondering because with the old ipfw implementation handling of both IP versions is separated into two different firewall programs, ipfw and ip6fw both being configured using distinct rc.conf variables. If IPv6 filtering is not supported by ipfw3 would it be better to configure a custom kernel without the INET6 option instead of using ip6fw alongside with ipfw3? Best Regards Matthias
initrd problem after upgrading from master
Hi, I updated my system to the latest master, today (26b5dbf28e0ee0c0...) and am having problems booting from initrd image using an encrypted ssd, now. The loader says it can't find /kernel/initrd. Also, I have regenerated initrd with make rescue, but this didn't help in resolving the issue. Here is the content of my loader.conf: autoboot_delay=3 loader_color=yes dm_load=yes dm_target_crypt_load=yes initrd.img_load=yes initrd.img_type=md_image vfs.root.mountfrom=ufs:md0s0 vfs.root.realroot=crypt:hammer:/dev/serno/number.s1d:root snd_hda_load=yes Do I have to issue an additional step apart from doing a make installkernel make installworld make upgrade? Regards Matthias
Re: Query about dragonflybsd LiveCD
Please be sure to use the right dev name! You might have a look into /var/log/messages to get the right name after connecting the stick. On 07.07.2015 11:59, Matthias Play wrote: You might als download dfly-x86_64-4.2.1_REL.img.bz2 instead and put it on an USB stick for booting. Issue the command: bzcat dfly-x86_64-4.2.1_REL.img.bz2 | dd of=USB dev bs=1M and replace usb dev with the stick's dev name (something like /dev/daX with X=8). After that you should be able to boot from the stick. On 07.07.2015 10:17, Bret Busby wrote: Hello. At the web page at http://www.dragonflybsd.org/download/ is stated that the iso for dragonflybsd is a LiveCD, and the image is at http://mirror-master.dragonflybsd.org/iso-images/dfly-x86_64-4.2.1_REL.iso.bz2 which is a bz2 file of about 200MB, which I have downloaded. I did not see any instructions for dealing with this file, so, is it a self-extracting iso file, such that the bz2 file is to be written to a disc, and then the disc with the bz2 file, used for booting, or, is the file required to be extracted to a directory which is then to be written to the disc for booting? Also, with the previous references to the mate GUI, does the iso image come with mate incorporated into it, so that a user can boot the LiveCD and thence examine the mate GUI? Thank you in anticipation.
Re: Query about dragonflybsd LiveCD
You might als download dfly-x86_64-4.2.1_REL.img.bz2 instead and put it on an USB stick for booting. Issue the command: bzcat dfly-x86_64-4.2.1_REL.img.bz2 | dd of=USB dev bs=1M and replace usb dev with the stick's dev name (something like /dev/daX with X=8). After that you should be able to boot from the stick. On 07.07.2015 10:17, Bret Busby wrote: Hello. At the web page at http://www.dragonflybsd.org/download/ is stated that the iso for dragonflybsd is a LiveCD, and the image is at http://mirror-master.dragonflybsd.org/iso-images/dfly-x86_64-4.2.1_REL.iso.bz2 which is a bz2 file of about 200MB, which I have downloaded. I did not see any instructions for dealing with this file, so, is it a self-extracting iso file, such that the bz2 file is to be written to a disc, and then the disc with the bz2 file, used for booting, or, is the file required to be extracted to a directory which is then to be written to the disc for booting? Also, with the previous references to the mate GUI, does the iso image come with mate incorporated into it, so that a user can boot the LiveCD and thence examine the mate GUI? Thank you in anticipation.
Re: DragonFly 4.2.1 released
Be sure you have initrd.img.gz in /boot/kernel and /boot is on an UFS formatted partition. If not, try regenerating it with mkinitrd -b bootdir. Regards, Matthias On 02.07.2015 23:12, Curtis Gagliardi wrote: I just went through this process upgrading from 4.2.0 to 4.2.1 and now I'm getting errors with mounting the filesystem. A potential key piece of info is that I'm dual booting with debian, which is the first OS on the disk and has an encrypted lvm group. I'm seeing dm_target_[error/zero/crypt]: successfully initialized and then an attempt to mount ufs:msdos0 with an error about there being so such disk. Full errors here: http://i.imgur.com/QFxfOHt.jpg Anyone know what's going wrong, or how I can start going about debugging this? My initial guess is that it's trying to mount my debian partition. I used the ? command to list the disk devices, and tried each one by running hammer:device, but they all failed. Thanks, Curtis On Wed, Jul 1, 2015, at 06:31 PM, Justin Sherrill wrote: Yep, it's a 0.0.1 release, to fix a bug in i915 support. If you aren't running an Intel chipset, there's no urgent need to update. You do not have to be at 4.2.0 to update to 4.2.1; for all intents and purposes, 4.2.1 can be treated the same as 4.2.0. I've uploaded the new images and they should be showing up on mirrors shortly. For those already running DragonFly, the same update methods as before apply: cd /usr/src git fetch origin git branch DragonFly_RELEASE_4_2 origin/DragonFly_RELEASE_4_2 git checkout DragonFly_RELEASE_4_2 git pull And then rebuild: make buildworld make buildkernel make installkernel make installworld make upgrade Any existing 4.2 packages you have installed are completely compatible.
Aw: AW: Re: Development branch
You may also use the source repo. You should be able to check out tagged revision v4.0.5 before you start the build: git checkout v4.0.5. In case you want to follow a release branch, let us say 4.0, you can check this revision out with: git checkout DragonFly_RELEASE_4_0 and update your local copy as usual with git pull. The master branch really is the devoted to DragonFly development and that is you get every developer commit whereas the release branches only contain revisions of DragonFly releases. Gesendet:Donnerstag, 11. Juni 2015 um 17:40 Uhr Von:nans_nans1 nans_na...@yahoo.de An:Matthias Sund matthias_p...@gmx.net, users@dragonflybsd.org Betreff:Aw: AW: Re: Development branch Ok. But sorry for my further question: And how can i update to 4.0.5 if i running 4.0.4? What is here difference? Von Samsung Mobile gesendet Ursprngliche Nachricht Von: Matthias Sund matthias_p...@gmx.net Datum: 11.06.2015 14:37 (GMT+01:00) An: nans_na...@yahoo.de Betreff: Aw: AW: Re: Development branch Hi, you need to check out the master branch from DragonFly source repo: cd /usr; make src-create make src-checkout You should have checked out master branch after that. You might check that with: cd /usr/src; git branch which should write * master to the terminal. To actually upgrade the system, you need to issue the following steps with optional steps in paranthesis: 1. cd /usr/src 2. (git pull) updates the sources to the latest devel version 3. make buildworld make buildkernel 4. make installkernel make installworld make upgrade 5. reboot In step 3, you may use make -j[number of available CPU cores] ... to speed up the build processes. Regards Matthias Gesendet:Donnerstag, 11. Juni 2015 um 13:49 Uhr Von:nans_nans1 nans_na...@yahoo.de An:John Marino dragonfly...@marino.st, users@dragonflybsd.org Betreff:AW: Re: Development branch Thank you. But, please, give me instructions what i have to do in detail. Von Samsung Mobile gesendet Ursprngliche Nachricht Von: John Marino dragonfly...@marino.st Datum: 11.06.2015 13:25 (GMT+01:00) An: nans_na...@yahoo.de,users@dragonflybsd.org Betreff: Re: Development branch On 6/11/2015 13:18, nans_na...@yahoo.de wrote: Hi is it possible to stay in development branch for every update forever? If yes, how can i do this? Build from source and have git set to the master branch. By definition, it is always development. It got changed to version 4.3 (from 4.1) this morning.
Re: domain name / host name
(Trying to answer the original question a bit more broadly). This is not a DragonFly specific distinction, it exists in the same way in all Unix-ish systems I know of; some are just more defaulty about it. OS X, for example, asks you for a computer name during setup (which it mangles to make a hostname without spaces), but just defaults the domain name to .local. Traditionally, the hostname represents the identity of a machine, and the domain the network it belongs to. DNS and other mechanisms help resolving *the combination of both* to a routable (IP) address, but they build on top of the original concept and are not a prerequisite. The hostname in this sense is a variable in memory, you can manipulate and query it using the hostname command. The domain is set in /etc/hosts and read from there. There is a bit of divergence in what to use as the hostname. Hence, more terminology was created: When you enter thehost as hostname, and example.org as domain, your box has a Fully Qualified Domain Name (FQDN) of thehost.example.org. When you type hostname on the shell, you get thehost, hostname -f gives thehost.example.org. However, in many installations, people just set the FQDN as the hostname. For example, Chef used to have some weird edge cases if you don't do this. In that case, hostname would give the FQDN, as would hostname -f. Since you sometimes still want just the host part, there is hostname -s which always only gives the part before the first dot. Of you want to specifically talk about this part, it's called the short hostname. Put together, FQDN = short hostname + domain; hostname = either short hostname or FQDN. All this is internal to the node; it does not matter whether the domain exists or is resolvable. It's purely self-identification up to this point. However, many applications use this identity in contexts where it is communicated to others. One example is sending email; unless specifically configured otherwise, sending email as user emily on our example host will generate it with a sender address of emily@thehost.example. org. The daily scripts just happen to be something that sends email from a vanilla DragonFly box (but normally only locally). Spam filters frequently back-check if the name you say you are sending from resolves back to you (in various ways); so if your FQDN is not resolvable you are very likely to have this email classified as spam. Another example are local network sharing/discovery protocols; SMB for windows networks and Bonjour for Mac networks being the most notable. Both include (among other things) mechanisms to a) enumerate all participating hostnames in the local network and b) resolve those names back to IP addresses. On DragonFly, both require additional software (samba for SMB, avahi for Bonjour); but when you install and enable them they use the hostname. I hope that makes it more clear what these are about? Matthias On Fri, May 29, 2015, 19:22 Christoph Harder shadow...@arcor.de wrote: In the setup one may provide a hostname and domain for the computer. How and where is this information exactly used? Hostname is pretty self-explaining, though I'm not sure, can it be used to access it like in a local windows network e.g. \\mycomputer\shareddisk\ e.g. after setting up an NFS share? And domain, is a real domain name required/recommend? One that is resolvable through DNS? And what happens if there are multiple domains hosted on a single server'?
Re: USB scanner device support
Hi Predrac, thanks for your extensive explanation. My problem was that I did not install SANE because I thought the device driver needs to be existent in the kernel. After installing SANE the scanner is functional and I can use xsane for image scans. Regards Matthias On 19.05.2015 00:59, Predrag Punosevac wrote: Matthias Play wrote: Hi, has scanner support been dropped in the past? My scanner is not recognized, the kernel message is: May 17 14:26:55 MisterX kernel: ugen3.2: Canon at usbus3 May 17 14:26:55 MisterX root: Unknown USB device: vendor 0x04a9 product 0x2206 bus uhub3 I have a Canon N 650U scanner and want to be able to scan using my DragonFly installation. What are my options to get it running? Regards Matthias Hi Matthias, I don't use DF on the desktop but I just checked /etc/sane.d on my OpenBSD 5.7 with the following version of sane-backend predrag@oko$ pkg_info sane-backends Information for inst:sane-backends-1.0.24p2 Quick greping reveals that the specific product is missing. predrag@oko$ grep 0x04a9 * canon630u.conf:usb 0x04a9 0x2204 canon_dr.conf:usb 0x04a9 0x1601 canon_dr.conf:usb 0x04a9 0x1602 canon_dr.conf:usb 0x04a9 0x1603 canon_dr.conf:usb 0x04a9 0x1604 canon_dr.conf:usb 0x04a9 0x1606 canon_dr.conf:usb 0x04a9 0x1607 canon_dr.conf:usb 0x04a9 0x1608 canon_dr.conf:usb 0x04a9 0x1609 canon_dr.conf:usb 0x04a9 0x160a canon_dr.conf:usb 0x04a9 0x160b canon_dr.conf:usb 0x04a9 0x genesys.conf:usb 0x04a9 0x2213 genesys.conf:usb 0x04a9 0x221c genesys.conf:usb 0x04a9 0x2228 genesys.conf:usb 0x04a9 0x1904 genesys.conf:usb 0x04a9 0x1909 genesys.conf:usb 0x04a9 0x1905 genesys.conf:usb 0x04a9 0x1906 genesys.conf:usb 0x04a9 0x1907 genesys.conf:usb 0x04a9 0x190a genesys.conf:usb 0x04a9 0x1906 If I was you I would try editing canon630u.conf which is on your machine probably located in /usr/local/etc/sane.d/ usb 0x04a9 0x2206 After you add it reboot the machine and give it a try. If that doesn't work try editing canon_dr.conf with the same line and reboot. I checked sane-backends page and your scanner indeed appears to be supported http://www.sane-project.org/sane-backends.html#S-CANON-DR IIRC that is one of those cheap scanners without power supply just using USB to power up. Those are really iffy on any system. For the record I have been using Epson Perfection 1670 which requires firmware for the past 8 years on OpenBSD and I couldn't be happier. Best, Predrag
USB scanner device support
Hi, has scanner support been dropped in the past? My scanner is not recognized, the kernel message is: May 17 14:26:55 MisterX kernel: ugen3.2: Canon at usbus3 May 17 14:26:55 MisterX root: Unknown USB device: vendor 0x04a9 product 0x2206 bus uhub3 I have a Canon N 650U scanner and want to be able to scan using my DragonFly installation. What are my options to get it running? Regards Matthias
Re: OT: can someone help me with an open-source CAD program?
Hi Pierre, for the GUI part I would recommend Ultimate++ (www.ultimatepp.org) that is a C++ framework which allows to build GUI apps quite fast. It is also a nice library if you intend to use a SQL db. The documentation is not that comprehensive, but one gets responses to questions very quickly in the forums of the project. Regards Matthias On 16.02.2015 18:53, Pierre Abbat wrote: I've been working on a surveying CAD program for several years and it's nearing the point where I'd like others to hack on it. I need help with the file format, import and export, and GUI. It runs on DragonFly (my build system is hosed, but I just ran it on leaf) and Linux. If you can help, or know someone who can, let me know. Pierre
Re: SMP Firewall
Yes. There are two resources where it will get you farther: interrupts and states. You will be pushing a lot of packets (=lots of interrupts to get them off the NIC) in this setup, with a non-SMP firewall one core has to deal with all of them. DragonFly takes great care to spread the processing as much as possible. With web requests you will probably have many, relatively low-volume connections, that amounts to a lot of state the firewall/LB has to take care of (which backend does this packet go to). SMP helps with the bookkeeping. Both of these do not degrade gracefully in my experience: you hit the limit and performance falls off a cliff. If you can, run some load tests to know when that happens. /mr On Sun, Feb 1, 2015, 03:58 Jeremy dyr...@gmail.com wrote: Does SMP matter to a firewall? For example: IF I was using one machine to load balance to 3 other web servers. Would SMP affect how it handles traffic? -Jeremy
Re: NAS setup recommendations
Hi Mehmet, thanks for your quick reply and suggestions! I would prefer to also always have a 1:1 copy of the OS part, because I also want to avoid exactly what you described i.e., do an extra install and configuration. It would maybe be sufficient to only do a cpdup from time to time to keep the OS copy up to date. I think this could be a good compromise. Regards Matthias Am 21.01.2015 um 19:17 schrieb Mehmet Erol Sanliturk: On Wed, Jan 21, 2015 at 9:22 AM, Matthias Play matthias_p...@gmx.net mailto:matthias_p...@gmx.net wrote: Hi, since I am about to setup a NAS box for private and business usage (it will also be providing additional services), I would like to ask for recommendations on what I might use. Maybe somebody already had similar intentions and realized that setup. The hardware is already in place and it basically contains HW proposed here (except for the HDD configuration and wih an additional 480G SSD for swapcache): http://blog.brianmoses.net/__2015/01/diy-nas-2015-edition.__html http://blog.brianmoses.net/2015/01/diy-nas-2015-edition.html. Ideally I would like to setup a RAID1 with a small boot partition and encrypting the rest just like it is demonstrated in encrypted_root.sh in the /usr/share/examples/rcconfig dir of a DragonFly installation. My high level requirements are: 1. encrypt as much as possible i.e., also / if possible 2. implement resilience to minimize down times 3. using hammer as the predominant file system to allow for fine grained fs snapshots In order to also support RAID1 I tried to set up vinum before encrypting the vinum volume with dm. But this has not worked out, maybe I did something wrong in configuring the loader variables for initrd. I read that it was planned to add mirror support in dm, but it has not been implemented so far. Maybe the reason is that with hammer there is already a descent mirror solution available. Although I think hammer mirror does not suit me 100% because of the expected down time in case of the master drive failing i.e., on would have to convert slave pfs to master pfs or cpdup the slave pfs contents to newly setup master pfs. Is there a configuration I might use to realize my three requirements above? Would using LVM be a viable alternative? Regards Matthas My suggestion may be the following : Use an independent disk for operating system . Format other disks for storing data . Mount them as independent drives ( drive names other than drive name of the operating system disk ) . Define an owner for the data disks ( the same user for all of them , not root ) Later on , when it becomes necessary to install a new operating system version , install it onto a new disk with the existing owner user name of the data disks in another computer . After verifying that the new operating system is working as expected , replace old operating system disk with the new one . In that way , data disks are recognized as they are because owner name is not changed . Installation of new operating system does not affect data disks where it is assumed that new version does not scratch existing data disks , i.e. disk format is not changed . With this practice , only down time required is to replace time of the new operating system disk . No loss of data , no complete copy onto other disks because operating system is in an independent disk . Thank you very much . Mehmet Erol Sanliturk
NAS setup recommendations
Hi, since I am about to setup a NAS box for private and business usage (it will also be providing additional services), I would like to ask for recommendations on what I might use. Maybe somebody already had similar intentions and realized that setup. The hardware is already in place and it basically contains HW proposed here (except for the HDD configuration and wih an additional 480G SSD for swapcache): http://blog.brianmoses.net/2015/01/diy-nas-2015-edition.html. Ideally I would like to setup a RAID1 with a small boot partition and encrypting the rest just like it is demonstrated in encrypted_root.sh in the /usr/share/examples/rcconfig dir of a DragonFly installation. My high level requirements are: 1. encrypt as much as possible i.e., also / if possible 2. implement resilience to minimize down times 3. using hammer as the predominant file system to allow for fine grained fs snapshots In order to also support RAID1 I tried to set up vinum before encrypting the vinum volume with dm. But this has not worked out, maybe I did something wrong in configuring the loader variables for initrd. I read that it was planned to add mirror support in dm, but it has not been implemented so far. Maybe the reason is that with hammer there is already a descent mirror solution available. Although I think hammer mirror does not suit me 100% because of the expected down time in case of the master drive failing i.e., on would have to convert slave pfs to master pfs or cpdup the slave pfs contents to newly setup master pfs. Is there a configuration I might use to realize my three requirements above? Would using LVM be a viable alternative? Regards Matthas
Most stable WiFi driver/chipsets?
Hey, which WLAN driver is currently the most stable with current chipsets? I.e. which chipsets/vendors should I be looking at buying? Background: my home router is a APU[0], currently running OpenBSD but I am looking at converting it to DragonFly. Additionally, I want to get rid of the external WLAN AP I am using right now and have the box serve this directly, so I can e.g. have a guest WiFi that is firewalled off from the internal network. Any input and experiences appreciated! From what I know so far I think I'd aim for Atheros since that seems to work pretty much everywhere, but I don't know what the status of the DragonFly driver is. Cheers, MR [0] http://www.pcengines.ch/apu1d.htm
Re: Cannot install xorg on dfly-x86_64-4.0.1_REL
I see this quite a lot on Linux; space taken up on disk (df) but not visible (du) is usually held by a deleted file that is still opened by some process. You can see these with lsof, and a reboot does indeed clear this. If you don't want to reboot, track down the process and kill it, this also releases the handle. /mr On Wed, Dec 31, 2014, 15:12 Aero 9000 mbg11...@gmail.com wrote: Quite odd, the 109M gap between the results of du -sh . (in /var) and du -h, disappeared after a reboot. I can't recall ever having had something like this in Linux, so in DragonFlyBSD I hadn't expected this behaviour at all. Like a dejavu of Windows -- to make it work, reboot.
Re: two crashes
Yes, you are right. Adding a few sentences to the documentation e.g., the handbook would be nice in order to make it easier for new users installing RC versions or master in release preparation times. Regards Matthias John Marino dragonfly...@marino.st schrieb: On 11/7/2014 23:07, Matthias wrote: Thanks Matt for your explanation. This also means that this ABI pkg problem always occurrs during release time for a short time period for RC versions and the master branch? what problem ? 3.6 = 3.6 3.7 = 3.8 3.8 = 3.8 3.9 = 3.10 3.10 = 3.10 3.11 = 3.12 4.0 = 4.0 4.1 = 4.2 4.2 = 4.2 That's how it's supposed to be, specifically to avoid busting all packages the second the release is branched. had 4.0 packages been in place before RC was announced, nobody would have noticed. There's no problem except people didn't know how to tell pkg to use packages with different but compatible ABI as a workaround. John
Re: Image corruption with owncloud + nginx + php-fpm
333
Request for MFC: OpenSSL update for CVE-2014-0160
Hi, I saw that Peter Avalos already updated OpenSSL on master, thank you for that! Could this also be backported to the 3.6 branch? It is also affected. More info: http://heartbleed.com/ http://filippo.io/Heartbleed/#www.dragonflybsd.org Thank you! Matthias
Re: ext2 or ext3 usage
Hi, this may have changed somewhat recently, but last time I attempted to use ext2/3 under *BSD (which AFAIK mostly share the driver) the situation was: * ext3 won't work * ext2 created with the default mke2fs settings won't work (unsupported options) * very basic, oldskool ext2 might work, but I haven't found the right settings yet I didn't pursue it further at the time. Best, Matthias On Tue, Dec 24, 2013 at 6:32 AM, Andrey Oktyabrskiy a...@bestmx.net wrote: Good day. Is it possible to mount ext2 or (better) ext3 FS in r/w mode under dfly? I've tried to mount both with the same result: $ sudo mount_ext2fs /dev/da0s2 /UNIT mount_ext2fs: /dev/da0s2: Invalid argument $ sudo fdisk /dev/da0 *** Working on device /dev/da0 *** parameters extracted from device are: cylinders=121126 heads=256 sectors/track=63 (16128 blks/cyl) Figures below won't work with BIOS for partitions not in cyl 1 parameters to be used for BIOS calculations are: cylinders=121126 heads=256 sectors/track=63 (16128 blks/cyl) Media sector size is 512 Warning: BIOS sector numbering starts with sector 1 Information from DOS bootblock is: The data for partition 1 is: sysid 165,(DragonFly/FreeBSD/NetBSD/386BSD) start 2079, size 134204931 (65529 Meg), flag 80 (active) beg: cyl 0/ head 33/ sector 1; end: cyl 1023/ head 255/ sector 63 The data for partition 2 is: sysid 131,(Linux filesystem) start 134219776, size 1819303936 (888332 Meg), flag 0 beg: cyl 1023/ head 254/ sector 63; end: cyl 1023/ head 254/ sector 63 The data for partition 3 is: UNUSED The data for partition 4 is: UNUSED What I do wrong?
Chaos Communication Congress 2012
Hi everybody, this is my annual reminder about the upcoming 29C3, the Chaos Communication Congress, a hacker gathering in Germany. For the first time the Congress takes place in the Congress Center Hamburg (CCH). The CCH is much larger than the BCC in Berlin, that means enough space for everyone! This also includes tickets. You can pre-order your ticket here: https://presale.events.ccc.de/ Due to the large space in the CCH there will be the possibility to set up an assembly. An Assembly is similar to a Village on the German/Dutch/American Hacker Camps. Maybe someone organizes one for DragonFly BSD?!? I glad to help. More information: http://events.ccc.de/2012/10/22/assemblies/ I'll be there for sure! What about the others? Cheers, Matthias
