Here is an interesting tutorial with suggestions on what to do when things go wrong with FreeBSD.
FreeBSD Ports: What To Do When Things Go Wrong
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Maybe you should check /usr/ports/sysutils/portdowngrade.
From the pkg description:
Portdowngrade helps to downgrade FreeBSD ports by analyzing the history of commits to the port and presenting the user the list of changes. By selecting one, the port can be set back to a previous version easily.
Oh, and if you dont want to build the INDEX yourself, you can just do:
cd /usr/ports && make fetchindex
Maybe you should check /usr/ports/sysutils/portdowngrade.
Tried that. The port was broken
The example from the pkg_which manpage helps find suspicious packages:
Get a list of files under /usr/local and /usr/X11R6 that do not belong to any package:
find /usr/local /usr/X11R6 -type f | xargs pkg_which -v | fgrep ‘?’
or try:
find /usr/local /usr/X11R6 -type f -print0 | xargs -0 pkg_which -v | fgrep ‘?’ | cut -f 1 -d :
I have problems with freebsd 4.10 as well. Give 5.2.1. a shot, it worked for me. I also got to say that article is awesome, helped me out.
Yeah, 4.10 has turned out to be a real stinker, and I only keep a copy around to more easilly set up a DragonFly system (I’m too lazy to manually install it ;^)
At any rate, FreeBSD 5.2.1 isn’t that bad, as long as you’re not using it for anything important (same with DragonFly).
Daan,
If you want to use it for desktop use I really recommend going for 5.2.1 instead of 4.10. 5.2.1 is quite stable and close to 5.3 which will replace 4.x stable branch in the near future.
Searching/using FBSD mailing lists will most probably help you alot more than posting your problem here.
pts
About your IDE problem – I’d suggest writing to [email protected], there’s bound to be someone that can help you there.
For the sound card, I’m pretty sure you’re loading the wrong kernel module, as sbc is for ISA sound cards. If you have any kind of modern system, you do not have an ISA sound card. See http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/sound-set… for more information on setting up sound, and I suggest you follow their advice and ‘kldload snd’ (FreeBSD 4.x) or ‘kldload snd_driver’ (FreeBSD 5.x) to load all drivers and see if yours is recognized.
For the sound card problem, disabling the “Plug and play OS” option in the BIOS would be another solution, unless you’ve already done it of course.
That is most likely a hardware problem and not the fualt of freebsd. Most likely bad cable, controller, or drive.
BTW I’ve seen stuff labeled SB16 compatable, even when it is not…
It looks like it never fully installed too… with that linking error…
Not sure of that, for my sisters box, which has a old isa sound blaster card in it, I just put pcm in the kernel and it worked fine.
I have problems with freebsd 4.10 as well. Give 5.2.1. a shot, it worked for me. I also got to say that article is awesome, helped me out.
FreeBSD 5 is slow as molasses (yes, yes, debugging is turned off) and is pretty unstable.
FreeBSD 5.2.1 is working flawlessly as my webserver and mailserver. I haven’t run into any major problems. I even decided to live on the wild side and *gasp* install Apache2 for my web server! While I prefer Linux as a desktop system (ALSA, better support for desktop hardware), I find that FreeBSD makes setting up a server so simple that it is hard to resist. The “Power to Serve” indeed.
The combo of cvsup + pkgdb -F + portupgrade -a really helps keep things nicely up-to-date, which gives me a nice fuzzy feeling.
I remember when FreeBSD used to seem so mysterious back in 1999 when I bought a copy of 3.3 in a Chapter’s bookstore back in late 1999 or early 2000. I reckoned I would be joining the UNIX cult for real if I dove into the murky depths of the BSD system with its C Shell and its BSD-style init system… it was new and I didn’t know much about it.
How I miss those days…
Depends on what you have for your desktop, I’ll think you’ll find most of the time is makes little difference.
I’ve personally never had a problem with freebsd and desktop hardware.
whats netcraft got to do with an ide and soundcard problem
why because the debug message said something about an uptime?
LOL
A couple of things with respect to your write errors, providing the following:
1) Same hardware (controller) for 5 years, is it a new drive w/old controller (more info needed)? Does the speed of the controller match the speed of the drive, for example ATA133 to ATA 133, if not your going to have to use ‘atacontrol’. It could also be some of the following:
2) Bad cable
3) Bad controller
4) Bad drive.
And I know most people use the hadware issue first and foremost, however the errors your getting are idicitive of hardware failure.
Though the article might not be written too well, it does give handy information. For example, once I was using portupgrade, which is a Ruby program. Portupgrade went to upgrade Ruby, yet in the new version some include-files, essential for Ruby, were moved to a separate package. The old Portupgrade package of course didn’t depend on that, so it started to upgrade Ruby, after which Portupgrade started giving error messages.
Luckily, the binary package of Ruby on the FreeBSD server was still the old version, but if it hadn’t been, I would have had a problem, as back then I didn’t know how to checkout an older version of the ports tree.
I think the article fails to mention one important thing, though, and that is that you should not only try out things first on another PC and make decent backups, but that you should also document what you do, by keeping notes or logs. In that way, when something goes wrong, you might be able to find the cause of any problem much easier.
Now some more information on my hardware problems:
My controller is a 5 years old <VIA 82C586 ATA33 controller>. It works flawlessly on Linux, Windows, NetBSD and OpenBSD, with UDMA, so I don’t see how it could be a hardware problem.
I have once written a bug report, got the answer that this was a not a bug and that I should simply disable DMA. That’s not the solution I would think of when they list the controller as fully supported! So I wrote to [email protected], no answer, filed another bug report, no answer.
The problem is not specific to 4.10, by the way. I tried 4.8, 4.9, 4.10, 5.1 and 5.2 and the problem appears in all of these releases. Additionally, it seems that, in the 4.x series, I can only disable DMA after the installation, as the bootloader used by the 4.x installer does not support the set command.
# kldload snd
# cat /dev/sndstat
FreeBSD Audio Driver (newpcm)
Installed devices:
pcm0: <CMI8330> at io 0x530 irq 11 drq 0 bufsz 4096 (1p/1r/0v channels duplex)
Looks great, but when I now try to actually play sound, some kind of trtrtrtrtrtrtrtrtrtr comes out of the speakers instead of the actual sound.
Oh, of course I disabled “PnP OS” in the BIOS, or NetBSD would not detect my network cards.
But okay, I’ll try the stable@ mailing list, so before I get any replies (or don’t 🙂 let’s don’t speak about my harware problems anymore, okay?
…Linux is slow as slug and is pretty unstable. :3 :3 :3
Har har. OK my comment was a bit provocative, but FreeBSD 5 is acknowledged by even FreeBSD developers to be quite a bit slower than the 4 branch. The network stack I think was reported to be about half the speed of 4.9/10, syscalls, disk IO, context switching, etc are slower. Mostly it is because of extra locking operations.
It is also reasonably unstable at the moment. Just look at the mailing lists if you want proof of that. Quite a lot of fundamental stuff is going on, along with the expected number of problems.
If FreeBSD 5 were fast and stable it would be released as a 5.3-STABLE branch.
… but FreeBSD 5 is acknowledged by even FreeBSD developers to be quite a bit slower than the 4 branch. The network stack I think was reported to be about half the speed of 4.9/10, syscalls, disk IO, context switching, etc are slower.
If that were indeed true then Linux 2.6 would be toast when 5 turns into 5-STABLE as tests (http://bulk.fefe.de/scalability/) indicated that 5-CURRENT (at that time 5-CURRENT was more like in between 5.1 and 5.2-RELEASE) was just a tad bit slower than Linux 2.6 on some sythetic benchmarks. 5.2 is quite a bit faster than 5.1 (which had it’s rough edges).
It is also reasonably unstable at the moment. Just look at the mailing lists if you want proof of that. Quite a lot of fundamental stuff is going on, along with the expected number of problems.
Don’t quite know where you get your info from but I’ve not seen any of this. Sure there is the odd setup/hardware that borks out during the development of 5-CURRENT but 5.2.1-RELEASE with ULE have been stable as hell for me on 8+ machines (firewall/router, laptop, desktop and servers). Using the ULE scheduler have meant no regressions in stability and increase in performance over the standard scheduler in 5.2.1-RELEASE. Remember, 5 is still a development branch (which you also stated) and having this stability then is very impressive. Perhaps we have the same view of what is unstable, that is the BSD standards of being rock solid. If that is so then yes FreeBSD-5 is a bit unstable but if you are using other OSes as standard then you’re way off. Fedora Core 2 have borked out a few times on a development machine that we use but FreeBSD 5.2.1 have been über-stable on that very same one under identical tests. Remember, Linux 2.6 is released and is as far as I know not considered a development version in Fedora Core 2.
If FreeBSD 5 were fast and stable it would be released as a 5.3-STABLE branch.
It’s getting there… pretty soon…
5.x slowness -> read this:
http://www.freebsd.org/releases/5.2.1R/early-adopter.html
“A certain amount of debugging and diagnostic code is still in place to help track down problems in FreeBSD 5.X’s new features. This may cause FreeBSD 5.X to perform more slowly than 4-STABLE.”
4.x and 5.x ATA problems, mostly with VIA chipsets – you may want to have a look at this thread:
http://www.freebsdforums.org/forums/showthread.php?s=&threadid=2225…
A bug that affected my HP 9510i with a VIA board (VT82C686B) was corrected because I gave myself to the trouble of helping the ATA maintainer (Soren Schmidt). Installed last CVS current snapshot, done CVS updates, and patched as he required. Now it’s solved and correction will probably (I hope) be included in the next STABLE version.
I had no reason to install -current, but I had no choice since a lot of bugfixes already had been made over the last STABLE and he’s only working over 5.x code now.
As you can see, helping them will help yourself.
Bruno
Linux 2.6 kernel is far from slow.
actually that is a problem with portupgrade, not the ports… this is becuase when something like that happens the port is bumped up a version… this means portupgrade will upgrade it and if a new dependency is found it will be installed…
not trying to be rude or any thing, I still suspect the controller as being the problem and not the code as I have had one on a old SBC and it worked fine… iirc
“If that were indeed true then Linux 2.6 would be toast when 5 turns into 5-STABLE as tests (http://bulk.fefe.de/scalability/) indicated that 5-CURRENT (at that time 5-CURRENT was more like in between 5.1 and 5.2-RELEASE) was just a tad bit slower than Linux 2.6 on some sythetic benchmarks. 5.2 is quite a bit faster than 5.1 (which had it’s rough edges).
”
stop turning this into yet another OS flamewar. the freebsd vm has took some good number of ideas from ingo’s scheduler in 2.6 and build upon it. linux and freebsd shares developers ideas and code. so less melo dramatism and talking on virtues is going to help better
I’ve been using FreeBSD from 3.xx series and it simply rules. After working with Linux for a while, it is like a breath of fresh air.
I have a mini-pc-server for my local network, with samba, NFS, httpd, mysql and friends.. it originally had FreeBSD 4.5-Stable, and after upgrading and updating the base system, ports, and other stuff, it’s still running, with FreeBSD-4.10 STABLE of course. And, of course, it still runs awesome!!!
Problems with ports??? Ok, i had one or two with weird ports, but with a little tweaking they’re working flawlessly.
BSDero
i kind of removed my whole /var/pkg(?) once so i lost the information about my old ports, any way to fix that automatically or should i reinstall some day?
If that were indeed true then Linux 2.6 would be toast when 5 turns into 5-STABLE as tests (http://bulk.fefe.de/scalability/) indicated that 5-CURRENT (at that time 5-CURRENT was more like in between 5.1 and 5.2-RELEASE) was just a tad bit slower than Linux 2.6 on some sythetic benchmarks. 5.2 is quite a bit faster than 5.1 (which had it’s rough edges).
Umm, that link shows 2.6 beating FreeBSD 5, so no I don’t think Linux 2.6 would be toast. Here is some light reading for you.
http://marc.theaimsgroup.com/?l=freebsd-amd64&m=108490922809652&w=2
http://groups.google.com.au/groups?hl=en&lr=&ie=UTF-8&threadm=10837…
http://marc.theaimsgroup.com/?l=freebsd-current&m=108817711204787&w…
http://marc.theaimsgroup.com/?l=freebsd-current&m=108777815720270&w…
http://groups.google.com.au/groups?hl=en&lr=&ie=UTF-8&threadm=10857…
Don’t quite know where you get your info from but I’ve not seen any of this. Sure there is the odd setup/hardware that borks out during the development of 5-CURRENT but 5.2.1-RELEASE with ULE have been stable as hell for me on 8+ machines (firewall/router, laptop, desktop and servers). Using the ULE scheduler have meant no regressions in stability and increase in performance over the standard scheduler in 5.2.1-RELEASE. Remember, 5 is still a development branch (which you also stated) and having this stability then is very impressive. Perhaps we have the same view of what is unstable, that is the BSD standards of being rock solid. If that is so then yes FreeBSD-5 is a bit unstable but if you are using other OSes as standard then you’re way off. Fedora Core 2 have borked out a few times on a development machine that we use but FreeBSD 5.2.1 have been über-stable on that very same one under identical tests. Remember, Linux 2.6 is released and is as far as I know not considered a development version in Fedora Core 2.
I get my info from the mailing lists. It isn’t exactly a secret, you know. And please PLEASE spare me your anecdotes about FreeBSD being stable for your use. I’m sorry, that doesn’t mean anything.
Linux has *never* crashed for me or corrupted a filesystem or been unstable unless I’ve been fooling around with the kernel or something. That absolutely doesn’t mean Linux has always been stable in that time – I know it hasn’t.
If you had the slightest knowledge of current FreeBSD development, you would know that 5 isn’t very stable at the moment. And your comment about the ULE scheduler just proves it. The ULE scheduler still has lots of problems and regressions vs the 4BSD one.