Re: Bind ntpd on certain interface?
The problem here is not the list attitude, but your silly That's right, I've already done it, I know, I know when somebody corrects you. That makes developers angry. Obviously something was wrong with your configs, and you think you know what, but don't. And that's worse than knowing you don't know. Then you did, what? Compete with developers for your truth? Lame. Nobody can force you, but I'll encorauge you stop whining because people are harsh at you: They know what they're talking about, so listen them before listening yourself. Regards, Dani Nice Daemon escribis: Can you please leave? Can you please force me? Honestly are you really that stupid to not understand when your welcome? No, I'm certainly not stupid. I'm just *re*acting (to remind you; in case you are actually able to *read*, you should already know it). People (Henning, Theo) started to bark at me when I asked for help. They didn't provide any help, they just needed someone to throw their words at. Seems like they have a severe need for psycho analysis (but hey, this is well-known throughout the net for Theo!). I don't think that this is normal behaviour, and I don't think that people appreciate it being treated like this. It seems (for years and years) that this is your (OpenBSD's developers/communities/whatever) attitude, so be it. But don't think that people being insulted will actually give donations to you or pay money to buy a CD/DVD set. They will (at max) use your software and never return anything back to you (the project) because they know, out of their own memories, because they read the list or because they read about this on other places, that you will insult them. You are the kids that nobody wants to play with. That nobody wants to fall in love with, that will die alone. Unloved. But it would be so easy to change: Just say 'hi!' instead of 'what do you motherfucking prick want?!'. :) Do you think anybody likes to help a prick like you? The OpenBSD mailing list is the only place I don't seem to be welcome. And guess what: I can live with it. Proudly. Joe -- :wq Claudio
Re: /altroot
On Sun, Aug 16, 2009 at 12:54 AM, Philip Guentherguent...@gmail.com wrote: On Sat, Aug 15, 2009 at 3:44 PM, 46254625...@gmail.com wrote: Is it correct string for /etc/fstab? /dev/wd2d /altroot ffs xx 0 0 Assuming /dev/wd2d is the correct partition, yes. (You're looking at the daily(8) manpage, right?) same device i am using here (a for root, b for swap, c entire disk, d for /altroot, and so on...) just two advices: (1) the /altroot filesystem must have the same size as the root (a) one (it may be slightly larger on architectures where root does not start at the beginning of the disk, like i386); (2) as Philip suggests you must read daily(8) to know how enabling back up of the root filesystem. my personal choice is setting ROOTBACKUP to 1 by hand and then run /etc/daily, this way /altroot will survive in case a mistake in the root filesystem remains unnoticed for more than twenty four hours. Should df display the /altroot? Only if you mount it yourself. It is not normally mounted and therefore does not appear in 'df' output. (No point in displaying it, what with it being a duplicate of the root filesystem) indeed, filesystems with xx as mount option are never mounted on boot. it is a backup filesystem -- automatically mounting it is a bad idea. cheers, igor.
08/11/09 install46 i386 freezes before boot: acpi?
Hello, Installed latest snapshot of mirror ftp://ftp.spline.de/pub/OpenBSD/ NameSizeLast Modified File:install46.iso 245002 KB 08/11/0914:22:00 I installed it on a fujitsu siemens amilo 1425 which has had the pleasure of having had the releases 3.9. 4.0, 4.1, 4.2, 4.3, 4.4 and 4.5 installed on it. All of them installed fine and without problems. The installation went fine. Then I rebooted and: www.aei.mpg.de/~pau/openbsd46.jpg (I couldn't get a better picture) ... it freezes. It stays like that and nothing else happens It must be acpi, but I cannot set it off, because I am not prompted with boot And I cannot switch it off in the bios. I say it must be ACPI because when I press the key that it's supposed to suspend the laptop, it switches off the screen. When I press the power key, it resumes to the screen you can see in the picture. Now what? Thanks, Pau
Re: 08/11/09 install46 i386 freezes before boot: acpi?
Pau wrote: ... The installation went fine. Then I rebooted and: www.aei.mpg.de/~pau/openbsd46.jpg (I couldn't get a better picture) bah. you were too lazy to type five very short lines of text? To save anyone else from clicking on the picture: it stops while loading /boot. ... it freezes. It stays like that and nothing else happens It must be acpi, no. but I cannot set it off, because I am not prompted with boot Not only has the kernel not loaded yet, the boot loader hasn't loaded yet. ACPI doesn't enter the picture until the kernel is loaded. And I cannot switch it off in the bios. I say it must be ACPI because when I press the key that it's supposed to suspend the laptop, it switches off the screen. When I press the power key, it resumes to the screen you can see in the picture. Your logic escapes me here. Before the OS has loaded, when the BIOS is in complete control still. the suspend key suspends the system. All is as it should be. /boot failed to load. Why? I don't know. Bad disk? funny disk layout? bad RAM/system? Corrupted file system? Boot your CD or floppy, at the boot prompt, enter boot hd0a:/bsd, that will probably work (depends on how hosed your 'a' partition is). If it does, copy over a new copy of /boot and reinstall the boot loader (faq14). I don't see anything that changed in /boot since 4.5, so I don't think you have a version issue, something went wrong in the install of /boot on your 4.6-cur install. This isn't to say that you DON'T also have an acpi issue, but you have to get the kernel loaded before we worry about that. Nick.
undeadly.org IPv6 reachability
Hello Misc, Since a few day, I can't connect to Undeadly.org over IPv6 (works well over v4). Is there any issue going on these days ? Thanks, Denis
Re: stop in /usr/ports/lang/ruby after make update on openBSD 4.5 i386
--- On Sun, 8/16/09, Jacob Meuser jake...@sdf.lonestar.org wrote: From: Jacob Meuser jake...@sdf.lonestar.org Subject: Re: stop in /usr/ports/lang/ruby after make update on openBSD 4.5 i386 To: misc@openbsd.org misc@openbsd.org Date: Sunday, August 16, 2009, 12:09 AM On Sat, Aug 15, 2009 at 01:36:49AM -0700, Robert Gilaard wrote: Hi guys, I'm trying to update my ports file based on the instructions at: http://www.openbsd101.com/updating.html and /lang/ruby stops with an error. I'm at the last part of step 2 of 3 - Updating and Building your Ports and have given the make update command in /usr/ports/lang/ruby directory. The interesting parts of the error log is this part: hecking whether to use xft... yes checking for X11/Xft/Xft.h... no /usr/ports/x11/tk/8.5/w-tk-8.5.6/tk8.5.6/generic/tk.h:78:29: X11/Xlib.h: No such file or directory Does anyone have a clue? Thanks in advanced. did you install xbase45 _and_ xshare45? -- I really can remember that. Is there an easy way to find out? I installed this test machine in may or something. Brgds Robert
lack of wordexp(3)
Hi, I was wondering whether there is a special reason why OpenBSD doesn't include the wordexp/wordfree functions. It seems like they are part of POSIX.1 and OpenBSD does feature the similar glob(3) function. Best regards, Jona -- Worse is better Richard P. Gabriel
Cleaning Up After Patching
OpenBSD Misc, I have installed OpenBSD 4.5 and applied the patches that have been issued, as per FAQ 10.15 - Applying patches in OpenBSD. My only question is, is there anything I need to do to clean up /usr/src after the patching and compiling is over? Also, is it recommended to keep /usr/src on a separate partition? Thank you. -- Okai
Re: 08/11/09 install46 i386 freezes before boot: acpi?
Sorry, I don't get it. Is it about insulting and discrediting or being patronizing? bah too lazy What's the problem? I am not a system administrator. If this mailing list is intended to be only for such, you should specify it. Even if it was. Neither the subject of the email, nor the content of the e-mail was in the style of HELP MY LAPPY DOESNT WORK THANKS HELP NEEDED I tried to be informative and accurate, at least as much as I could. I followed the specifications in http://www.openbsd.org/mail.html about Netiquette This is btw not the first time this happens in this mailing list. I would have sufficed with these two paragraphs here Boot your CD or floppy, at the boot prompt, enter boot hd0a:/bsd, that will probably work (depends on how hosed your 'a' partition is). If it does, copy over a new copy of /boot and reinstall the boot loader (faq14). I don't see anything that changed in /boot since 4.5, so I don't think you have a version issue, something went wrong in the install of /boot on your 4.6-cur install. In any case, thanks for _these_ two paragraphs. Pau 2009/8/16 Nick Holland n...@holland-consulting.net: Pau wrote: ... The installation went fine. Then I rebooted and: www.aei.mpg.de/~pau/openbsd46.jpg (I couldn't get a better picture) bah. you were too lazy to type five very short lines of text? To save anyone else from clicking on the picture: it stops while loading /boot. ... it freezes. It stays like that and nothing else happens It must be acpi, no. but I cannot set it off, because I am not prompted with boot Not only has the kernel not loaded yet, the boot loader hasn't loaded yet. ACPI doesn't enter the picture until the kernel is loaded. And I cannot switch it off in the bios. I say it must be ACPI because when I press the key that it's supposed to suspend the laptop, it switches off the screen. When I press the power key, it resumes to the screen you can see in the picture. Your logic escapes me here. Before the OS has loaded, when the BIOS is in complete control still. the suspend key suspends the system. All is as it should be. /boot failed to load. Why? I don't know. Bad disk? funny disk layout? bad RAM/system? Corrupted file system? Boot your CD or floppy, at the boot prompt, enter boot hd0a:/bsd, that will probably work (depends on how hosed your 'a' partition is). If it does, copy over a new copy of /boot and reinstall the boot loader (faq14). I don't see anything that changed in /boot since 4.5, so I don't think you have a version issue, something went wrong in the install of /boot on your 4.6-cur install. This isn't to say that you DON'T also have an acpi issue, but you have to get the kernel loaded before we worry about that. Nick.
Re: /altroot
On Sat, 15 Aug 2009, Philip Guenther wrote: Is it correct string for /etc/fstab? /dev/wd2d /altroot ffs xx 0 0 Assuming /dev/wd2d is the correct partition, yes. (You're looking at the daily(8) manpage, right?) Sure. Original mount string for '/altroot' there was '/dev/wd2d /altroot ffs rw,nodev,nosuid 1 2'. Should df display the /altroot? Only if you mount it yourself. It is not normally mounted and therefore does not appear in 'df' output. (No point in displaying it, what with it being a duplicate of the root filesystem) Ok. Thanks. I do not see any messages about backup. How do I check if backup really happen? -- 4625
Re: /altroot
On Sun, 16 Aug 2009, Igor Sobrado wrote: Is it correct string for /etc/fstab? /dev/wd2d /altroot ffs xx 0 0 same device i am using here (a for root, b for swap, c entire disk, d for /altroot, and so on...) just two advices: (1) the /altroot filesystem must have the same size He have. Filesystem SizeUsed Avail Capacity Mounted on /dev/wd2a 223M 58.3M153M28%/ /dev/wd2d 223M 58.3M153M28%/altroot Wow, now I'm sure - backup happens here. But silencely... as the root (a) one (it may be slightly larger on architectures where root does not start at the beginning of the disk, like i386); (2) as Philip suggests you must read daily(8) to know how enabling back up of the root filesystem. my personal choice is setting ROOTBACKUP to 1 by hand and then run /etc/daily, this way /altroot will survive in case a mistake in the root filesystem remains unnoticed for more than twenty four hours. Should df display the /altroot? Only if you mount it yourself. It is not normally mounted and therefore does not appear in 'df' output. (No point in displaying it, what with it being a duplicate of the root filesystem) indeed, filesystems with xx as mount option are never mounted on boot. it is a backup filesystem -- automatically mounting it is a bad idea. -- 4625
Re: 08/11/09 install46 i386 freezes before boot: acpi?
On Sun, Aug 16, 2009 at 09:40:54PM +0200, Pau wrote: Sorry, I don't get it. Is it about insulting and discrediting or being patronizing? bah too lazy What's the problem? I am not a system administrator. If this mailing list is intended to be only for such, you should specify it. I don't get it either. Why is it that only people who use gmail and hide behind lame names feel stepped on their toes because someone expresses his opinion on their low quality mailings full of assumptions that don't even make the slightest sense? Mommy, mommy, the evil OpenBSD people tried to help me but they didn't deliver the answer^Wconfirmation that I wanted on a golden platter. I am so hurt, my soul is shattered and my eyes are filled with tears :B4( Even if it was. Neither the subject of the email, nor the content of the e-mail was in the style of HELP MY LAPPY DOESNT WORK THANKS HELP NEEDED I tried to be informative and accurate, at least as much as I could. I followed the specifications in http://www.openbsd.org/mail.html about Netiquette Which part of Plain text, 72 characters per line escaped you? This is btw not the first time this happens in this mailing list. I would have sufficed with these two paragraphs here Yeah, teach those people who try to help you how to answer. After all you're paying them big monies and if they want their operating system to succeed, they better help you. Otherwise you're going to write bad comments about them on slashdot and tell everyone that Linux is much better, because it gives you blowjobs while starting udev! Boot your CD or floppy, at the boot prompt, enter boot hd0a:/bsd, that will probably work (depends on how hosed your 'a' partition is). If it does, copy over a new copy of /boot and reinstall the boot loader (faq14). I don't see anything that changed in /boot since 4.5, so I don't think you have a version issue, something went wrong in the install of /boot on your 4.6-cur install. In any case, thanks for _these_ two paragraphs. Pau What goes around, comes around. 2009/8/16 Nick Holland n...@holland-consulting.net: Pau wrote: ... The installation went fine. Then I rebooted and: www.aei.mpg.de/~pau/openbsd46.jpg (I couldn't get a better picture) bah. you were too lazy to type five very short lines of text? To save anyone else from clicking on the picture: it stops while loading /boot. ... it freezes. It stays like that and nothing else happens It must be acpi, no. but I cannot set it off, because I am not prompted with boot Not only has the kernel not loaded yet, the boot loader hasn't loaded yet. ACPI doesn't enter the picture until the kernel is loaded. And I cannot switch it off in the bios. I say it must be ACPI because when I press the key that it's supposed to suspend the laptop, it switches off the screen. When I press the power key, it resumes to the screen you can see in the picture. Your logic escapes me here. Before the OS has loaded, when the BIOS is in complete control still. the suspend key suspends the system. All is as it should be. /boot failed to load. Why? I don't know. Bad disk? funny disk layout? bad RAM/system? Corrupted file system? Boot your CD or floppy, at the boot prompt, enter boot hd0a:/bsd, that will probably work (depends on how hosed your 'a' partition is). If it does, copy over a new copy of /boot and reinstall the boot loader (faq14). I don't see anything that changed in /boot since 4.5, so I don't think you have a version issue, something went wrong in the install of /boot on your 4.6-cur install. This isn't to say that you DON'T also have an acpi issue, but you have to get the kernel loaded before we worry about that. Nick.
Virus Email.Phishing.Acc-1 gefunden
Sehr geehrte Damen und Herren, in dem E-Mail mit dem Betreff 'Dear Webmaster Email Account Owner' (gesendet am Sun, 16 Aug 2009 22:43:27 GMT) mit der angegebenen Absenderadresse 'WEBMASTER webnas...@tech.net' wurde der Virus 'Email.Phishing.Acc-1' gefunden. Aus diesem Grund wurde die E-Mail nicht zugestellt! Ihr Inode-Team -- Dear Ladies and Gentlemen, the mail with the Subject 'Dear Webmaster Email Account Owner' (sent on Sun, 16 Aug 2009 22:43:27 GMT) with the sender address specified as 'WEBMASTER webnas...@tech.net' contained a virus known as 'Email.Phishing.Acc-1'. Due to this reason the Mail has not been delivered! Your Inode-Team --- Headers of original mail follow: From owner-misc+m88...@openbsd.org Mon Aug 17 01:25:14 2009 X-Envelope-From: owner-misc+m88...@openbsd.org Received: from lists.openbsd.org ([192.43.244.163]:11757 helo=shear.ucar.edu) by smartmx-02.inode.at with esmtp (Exim 4.69) (envelope-from owner-misc+m88...@openbsd.org) id 1Mcp6A-0002pl-Lz for te...@lynix.net; Mon, 17 Aug 2009 01:25:14 +0200 Received: from openbsd.org (localhost.ucar.edu [127.0.0.1]) by shear.ucar.edu (8.14.3/8.14.3) with ESMTP id n7GNNGpW006732; Sun, 16 Aug 2009 17:23:16 -0600 (MDT) Received: from cvs.openbsd.org (cvs.openbsd.org [199.185.137.3]) by shear.ucar.edu (8.14.3/8.14.3) with ESMTP id n7GNJrEm022186 (version=TLSv1/SSLv3 cipher=DHE-DSS-AES256-SHA bits=256 verify=FAIL) for m...@shear.ucar.edu; Sun, 16 Aug 2009 17:19:53 -0600 (MDT) Received: from jamesv.securesites.net (jamesv.securesites.net [128.241.54.188]) by cvs.openbsd.org (8.14.3/8.12.1) with ESMTP id n7GNJpeQ020121 (version=TLSv1/SSLv3 cipher=DHE-DSS-AES256-SHA bits=256 verify=NO) for misc@openbsd.org; Sun, 16 Aug 2009 17:19:52 -0600 (MDT) Received: from jamesv.securesites.net (localhost [127.0.0.1]) by jamesv.securesites.net (8.13.6.20060614/8.13.6) with ESMTP id n7GMhRji017105 for misc@openbsd.org; Sun, 16 Aug 2009 22:43:27 GMT Received: (from w...@localhost) by jamesv.securesites.net (8.13.6.20060614/8.13.6/Submit) id n7GMhRs3017101; Sun, 16 Aug 2009 22:43:27 GMT Date: Sun, 16 Aug 2009 22:43:27 GMT Message-Id: 200908162243.n7gmhrs3017...@jamesv.securesites.net To: misc@openbsd.org Subject: Dear Webmaster Email Account Owner From: WEBMASTER webnas...@tech.net Reply-To: webmasterwebmail.accou...@gmail.com MIME-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Loop: misc@openbsd.org Precedence: list Sender: owner-m...@openbsd.org
Boxes hanging intermittently. Anybody seen such ?
Hello, During the past week two boxes two boxes on the same network have stopped responding, they carry OpenBSD 4.5 i386 and I have logged at every possible log to find out why this occurs however I havent been able to spot anything unusual. All of the sudden they just stop responding requests. What these freezes do have in common is that when the boxes are reached via the KVM they present the login screen, they allow text to be entered in the login field... but upon hitting enter for it to ask the password thats when it just hangs. Iam afraid this will keep on happening and I woudl like to know if anybody has experienced this before.. these have been perfectly working boxes and it would be just odd both would have the same problem in the same few days. Upon a reboot everything returns to normal. Thank you. Andres
Re: /altroot
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Sun, Aug 16, 2009 at 09:44:43PM +, 4625 spoke thusly: On Sat, 15 Aug 2009, Philip Guenther wrote: Is it correct string for /etc/fstab? /dev/wd2d /altroot ffs xx 0 0 Assuming /dev/wd2d is the correct partition, yes. (You're looking at the daily(8) manpage, right?) Sure. Original mount string for '/altroot' there was '/dev/wd2d /altroot ffs rw,nodev,nosuid 1 2'. Should df display the /altroot? Only if you mount it yourself. It is not normally mounted and therefore does not appear in 'df' output. (No point in displaying it, what with it being a duplicate of the root filesystem) Ok. Thanks. I do not see any messages about backup. How do I check if backup really happen? -- 4625 It's in daily output mailed to root's account. You got to have access to root's mail. I've got myself in the wheel group and have root's email sent to me instead. Here's some pertinent stuff in root's crontab: ROOTBACKUP=1# backup / to /altroot MAILTO=dennyboy # forward root's mail to me In /etc/mail/aliases, my name is added next to root: root: dennyboy In daily output to root, the section of the message pertaining to the backup should look something similar to this: Backing up root filesystem: copying /dev/rwd0a to /dev/rwd1a 131098+1 records in 131098+1 records out 1073955328 bytes transferred in 51.680 secs (20780520 bytes/sec) Naturally the 2nd line is dependent on your disk setup, fstab so forth as previously discussed in this thread. Denny White - -- === () ASCII ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments === GnuPG key : 0x1644E79A | http://wwwkeys.de.pgp.net Fingerprint: D0A9 AD44 1F10 E09E 0E67 EC25 CB44 F2E5 1644 E79A === iEYEARECAAYFAkqIh3AACgkQy0Ty5RZE55rNAACgn6nB3rWPk5A9Pdl6c3wloMHh 4iMAn26SIqS+5lFH3d3VrBFlDYAxSXQk =HLK7 -END PGP SIGNATURE-
Re: Cleaning Up After Patching
Okai Mood wrote: OpenBSD Misc, I have installed OpenBSD 4.5 and applied the patches that have been issued, as per FAQ 10.15 - Applying patches in OpenBSD. My only question is, is there anything I need to do to clean up /usr/src after the patching and compiling is over? nope. Any needed cleanup will be taken care of at the start of the next build cycle. (possible exception: the patch files themselves, but I really don't think they will be big enough to cause you any problems, and leaving them in place might help remind you what patches have been applied and which haven't.) Also, is it recommended to keep /usr/src on a separate partition? Certainly not a bad thing. If you look at the default install on a big disk for 4.6, you see the following partitions and how they are mounted: /dev/wd0a on / type ffs (rw, local) /dev/wd0k on /home type ffs (rw, local, nodev, nosuid) /dev/wd0d on /tmp type ffs (rw, local, nodev, nosuid) /dev/wd0f on /usr type ffs (rw, local, nodev) /dev/wd0g on /usr/X11R6 type ffs (rw, local, nodev) /dev/wd0h on /usr/local type ffs (rw, local, nodev) /dev/wd0j on /usr/obj type ffs (rw, local, nodev, nosuid) /dev/wd0i on /usr/src type ffs (rw, local, nodev, nosuid) /dev/wd0e on /var type ffs (rw, local, nodev, nosuid) In addition to some logistical benefit, there is a security benefit here. Only root has write access to anything in most of /usr, with the exceptions of /usr/src, /usr/obj. Those two directories can, by default, be written by anyone in the wsrc group. Note that those two directories are nosuid, which reduces some of the mischief someone in the wsrc group could get into. This keeps with the general theme of, directories where users can write should be nosuid, nodev, areas that have to be mounted to permit devices and setuid apps need to be not writable by non-root users. Nick.
Re: 08/11/09 install46 i386 freezes before boot: acpi?
Pau wrote: Sorry, I don't get it. I'd suggest you think really long and hard about which you prefer when you accidentally post a note that people who could help you find annoying: 1) They point out why your message was annoying, and go on to help you. 2) They ignore you. I almost never look at people's screen shots of problems, For some reason something in your note caught my attention (namely, things didn't seem to make sense), so against my better judgment, I looked. Upon seeing you felt it simpler to post a photo rather than typing in five really easy lines of text, I was pissed enough I nearly didn't respond. I did. My mistake. Please accept my apology for not having ignored you as I should have. I'll try to do better in the future. Nick.
Cursor blink frequency.
How to decrease cursor blink frequency? -- 4625
Re: yt: youtube download issue
Howdy? It would appear that this is back again, and since it has happened *while* running 4.5 Rel(i386), in which yt initially worked, I suspect YouTube has changed something. Can anyone else confirm/deny this conjecture? Thanks Dhu On Tue, 18 Dec 2007 20:18:26 +0530 Girish Venkatachalam girishvenkatacha...@gmail.com wrote: On Tue, Dec 18, 2007 at 06:33:07PM +1100, Chris wrote: yt is giving me the following error while trying to download - $ yt http://youtube.com/watch?v=huF2mrhTtCwfeature=dir $ Getting http://youtube.com/watch?v=huF2mrhTtCw ... /usr/local/bin/lua: /usr/local/bin/yt:42: assertion failed! stack traceback: [C]: in function 'assert' /usr/local/bin/yt:42: in main chunk [C]: ? [2]+ Exit 1 yt http://youtube.com/watch?v=huF2mrhTtCw I have yt-6 and lua-5.1.2p0 installed on 4.2. Any help would be much appreciated. Thanks. Problem is fixed in -current. -Girish
Re: /altroot
On Sun, 16 Aug 2009, Denny White wrote: Sure. Original mount string for '/altroot' there was '/dev/wd2d /altroot ffs rw,nodev,nosuid 1 2'. I do not see any messages about backup. How do I check if backup really happen? In daily output to root, the section of the message pertaining to the backup should look something similar to this: Backing up root filesystem: 131098+1 records in 131098+1 records out I see now. So, everything is fine! -- 4625
Re: Cleaning Up After Patching
On Sun, Aug 16, 2009 at 6:46 PM, Nick
Hollandn...@holland-consulting.net wrote:
Okai Mood wrote:
OpenBSD Misc,
I have installed OpenBSD 4.5 and applied the patches that have been
issued, as per FAQ 10.15 - Applying patches in OpenBSD. My only
question is, is there anything I need to do to clean up /usr/src after
the patching and compiling is over?
nope. Any needed cleanup will be taken care of at the start of the
next build cycle.
(possible exception: the patch files themselves, but I really don't
think they will be big enough to cause you any problems, and leaving
them in place might help remind you what patches have been applied
and which haven't.)
Also, is it recommended to keep
/usr/src on a separate partition?
Certainly not a bad thing.
If you look at the default install on a big disk for 4.6, you see
the following partitions and how they are mounted:
/dev/wd0a on / type ffs (rw, local)
/dev/wd0k on /home type ffs (rw, local, nodev, nosuid)
/dev/wd0d on /tmp type ffs (rw, local, nodev, nosuid)
/dev/wd0f on /usr type ffs (rw, local, nodev)
/dev/wd0g on /usr/X11R6 type ffs (rw, local, nodev)
/dev/wd0h on /usr/local type ffs (rw, local, nodev)
/dev/wd0j on /usr/obj type ffs (rw, local, nodev, nosuid)
/dev/wd0i on /usr/src type ffs (rw, local, nodev, nosuid)
/dev/wd0e on /var type ffs (rw, local, nodev, nosuid)
In addition to some logistical benefit, there is a security benefit
here. Only root has write access to anything in most of /usr, with the
exceptions of /usr/src, /usr/obj. Those two directories can, by
default, be written by anyone in the wsrc group. Note that those two
directories are nosuid, which reduces some of the mischief someone
in the wsrc group could get into. This keeps with the general theme
of, directories where users can write should be nosuid, nodev, areas
that have to be mounted to permit devices and setuid apps need to be
not writable by non-root users.
Good points. This brings up a question I have meant to ask. Since we
are giving sources their own mount point, wouldn't it makes sense to
have a different name for this mount point (other than /usrc/src) so
that both /usr/ports and /usr/xenocara can also reside there? As is,
with the layout the installer suggests/offers, you are left with
/usr/{ports,xenocara} in the /usr.
What I've done on my -current system, I have a /usr/osrc mount point
and soft-links for /usr/{ports,src,xenocara} into that mount point.
Same with object directories:
$ ls -l /usr/{obj,ports,src,xenocara,xobj}
lrwxr-xr-x 1 root wheel 9 Jul 4 13:05 /usr/obj - oobj/obj/
lrwxr-xr-x 1 root wheel 11 Jul 4 13:05 /usr/ports - osrc/ports/
lrwxr-xr-x 1 root wheel 9 Jul 4 13:04 /usr/src - osrc/src/
lrwxr-xr-x 1 root wheel 14 Jul 4 13:05 /usr/xenocara - osrc/xenocara/
lrwxr-xr-x 1 root wheel 10 Jul 4 13:05 /usr/xobj - oobj/xobj/
The only draw back with this scheme seems to be a daily security
warning about /usr/src being a link and having a different gid.
I suppose, one could have different mount points for each of the five
directories mentioned above, but that could be a bit overkill if the
soft-links accomplish the same goal(s).
--patrick
