OpenBSD Journal

Jem Matzan of Newsforge likes OpenBSD 3.6

Contributed by grey on from the third party endorsements dept.

Thanks to Mark Patterson and others for writing in with the following:

Here's a positive review of OpenBSD 3.6, even recommending getting the CDs His summary is "what strikes me most about OpenBSD in general is the professional manner in which it is developed and released."

The review may be found here.

(Comments are closed)

Comments

  1. By almeida (66.31.180.15) on

    Good review.

  2. By Anonymous Coward (65.42.15.242) on

    this is a good review except that he says "OpenBSD 3.6 shows steady improvement" which makes it sound like it sucked before or something which is hardly the case and he says "The quickest way to get OpenBSD 3.6 is through an FTP install...Every time I try this I have some kind of problem" and doesn't go into any detail about what his problems were. i have not installed OpenBSD anyway but via ftp and have never had any problems, not to mention that fact that OpenBSD's ftp install is considerably faster than other *nix's cd installs!

  3. By Anthony (68.145.111.152) on

    3.5 was disapointing for me. I couldn't read any of my old filesystems (had to boot up with 3.4 to copy everything over to the new drive), and I could crash it if I hit it with sufficient network traffic. I didn't have any problems reading my old filesystems after the upgrade to 3.6, and I've been unable to crash it with any amount of network traffic. I'm much more comfortable with this release than the last one. One gets the feeling it was mostly a bugfix release for anyone using the UP kernel. There's new features apart from SMP, but the stability is what I needed. 3.2-3.4 were great, and it looks like 3.6 is great. I'm hoping 3.5 was an aberration. :)

    Comments

    1. By Brad (216.138.200.42) brad at comstyle dot com on

      You make it sound as if no one else but yourself had issues with 3.5. Some people would completely disagree with you and say 3.5 was better because they've run into issues with 3.6. There are always issues with any release of any operating system.

      Comments

      1. By LJ (62.59.31.68) lj@2u2.nu on

        Or they could agree with him. Point is that you don't need to reason away a comment in such a way that it appears insignificant.. I would like to know which type of network adapter this user is using.

        Comments

        1. By Anthony (68.145.111.152) on

          dmesg

          That dmesg is obviously from 3.6, but the hardware hasn't changed since I ran 3.5. em0 was the one giving me trouble.

          Henning e-mailed me after I posted to misc@ about it. He said the issue was fixed in CURRENT, and suggested I upgrade. I wasn't comfortable doing that, but CURRENT has since become 3.6-release, and it is indeed fixed.

          With 3.5 a "ping -f" from my Linux box would bring down the OpenBSD machine in seconds. Sometimes to a debugger prompt, sometimes causing a spontaneous reboot. With 3.6 I left it on all night without noticing any ill effects. Thus, I a) consider it to have been a bug and b) I consider that bug fixed.

          Comments

          1. By Anonymous Coward (216.238.113.174) on

            Henning e-mailed me after I posted to misc@ about it.

            You can't comaplain about the level of support you received. :-)

            Comments

            1. By Anthony (68.145.111.152) on

              Reading my old filesystems would have been nice. I never got a response from anyone about that.

              Comments

              1. By Nick Holland (68.43.115.33) nick@holland-consulting.net on http://www.openbsd.org/faq/

                You don't provide enough details (not even platform!), but I'd suspect it is this (FAQ 4).

                3.5 was much more picky about this, that was deliberate. If this was your problem, your system was broke before, we just let you know in no uncertain terms this time. Not like I and other people hadn't been saying "This is NOT the right way to set up your disks" for years.

                Reminder: OpenBSD 3.5 was the first release to support booting on i386 with a root partition that was not fully within the first 8G of a disk. As long as the boot code was being worked on, lots of new features went in, including a check on installboot so it would not put a PBR where only the MBR should be.

                All things considered, considering how few computers (I can't think of any, actually) that the all-new boot code broke, it's really incredible. I did a huge amount of testing of that code before it was shown to the general public (one of the downsides of having a huge repository of OpenBSD-capable i386 systems!), and I can tell you that it basically worked from the first version Tom showed me. Incredible work by Tom Cosgrove, and great design from Toby Weingartner (which eliminated the translation issues that plagued the old boot loader when moving a hard disk from machine to machine.

                There is nothing which should have prevented 3.5 from reading a healthy 3.4 file system, I feel fairly confident in saying your machine was broke in some way.

                Comments

                1. By Anthony (68.145.111.152) on

                  "<i>There is nothing which should have prevented 3.5 from reading a healthy 3.4 file system, I feel fairly confident in saying your machine was broke in some way.</i>"

                  I'm disinclined to accept that explaination without further evidence, as it would imply that my old OpenBSD disks were broke, my USB disk was broken, and another unrelated hard drive was broken. That's not impossible, but it's unlikely enough that I'm not going to assume it's the case.

                  My pre-upgrade system had two hard disks. I couldn't read any filesystem on either with 3.5 (or 3.6), but the partitions appeared correct with fdisk. That was a fresh install from 3.4, and I told fdisk to use the whole disk for OpenBSD, so I think it's unlikely that I screwed up. Also, the system was (and is) unable to mount either a backup hard drive I move from place to place that's one giant ext2[1] partition or either of the filesystems on my USB key (one FAT32 and one ext2).

                  In practice, I don't care because my Mac and my Linux box can read these, and it's a lot easier to put drives in these machines (they're not in the crawlspace). However, even if these other OSes are creating screwed up filesystems or partitions (easily possible), there's not much point in having FAT32 or ext2 drivers if they can't read partitions created by other OSes.

                  When I get home I will post with error messages I get with the USB disk.

                  1-The only thing all of my *nixen can read.

                2. By Anthony (68.145.111.152) on

                  The problem I'm having now is a "device not configured" when attempting to mount /dev/sd0i (the ext2 filesystem) on the USB disk. It's possible it's a USB issue rather than a disk issue. It's also possible that I had this same problem the last time and simply remember incorrectly, and therefore my willingness to accept your theory has increased.

                  I remain less than thrilled at being unable to mount the filesystem, but I'm not sure it's not my fault and figuring this out will probably take longer than it's worth.

                  btw, the dmesg for the machine (same hardware, now 3.6) is higher up in the thread.

                  Comments

                  1. By Anonymous Coward (213.84.84.111) on

                    You demsg does not show a USB disk. A dmesg showing the disk and the output of disklabel sd0 (or whatever number the USB disk gets) might help.

          2. By Anonymous Coward (68.125.27.117) on

            "He said the issue was fixed in CURRENT, and suggested I upgrade. I wasn't comfortable doing that"

            Someone who knows what they are talking about suggests you do something and you don't do it?

            Let me repeat; -current is just as reliable as -stale and -release. -current is used on production systems with NO PROBLEMS!

            Comments

            1. By Anthony (68.145.111.152) on

              "Someone who knows what they are talking about suggests you do something and you don't do it?

              Let me repeat; -current is just as reliable as -stale and -release. -current is used on production systems with NO PROBLEMS!"

              My lack of skills, my risk: my decision.

              The risk of sticking with 3.5-stable was low: It didn't crash under normal loads. The external network was too slow for anyone to crash the machine, and all the users on the LAN could just as easily have yanked the power cable.

              The risk of upgrading to -current was high: Even if the code was bug free, I was still perfectly capable of fucking it up myself. From what Nick has said in this thread it seems likely that the problems I had upgrading 3.4 -> 3.5 were completely my fault, so this is clearly possible.

              I'm still new enough at this that I can't safely assume I won't screw up given the opportunity. Given that I was using Windows a year ago, with OpenBSD on my spare machine so I could poke around when I wanted to experiment, I think my progress has been satisfactory. However, I am no expert.

    2. By tedu (67.124.88.142) on

      did you tell anyone about this filesystem trouble of yours? because you're certainly the only one who had this problem. my 3.4 (no, wait, make that 3.0) filesystems worked just fine with 3.5. we don't fix problems we don't know about.

      Comments

      1. By Anthony (68.145.111.152) on

        I posted to misc@ and no one replied, but Nick posted in this thread. He thinks I had an already-broken setup, and I've got no reason to reject that at this point.

  4. By Anonymous Coward (66.91.134.210) on

    his idea of a portsupgrade application is pointless I realize the ports tree is "nice" but not part of what the openbsd team needs to concern themselves with.

    Comments

    1. By Brad (216.138.200.42) brad at comstyle dot com on

      If you've been paying attention to -current then you would know that a great deal of work has been done to deal with package upgrading. By the time 3.7 rolls around the improved package tools should be able to deal with upgrading of packages and be way better than the nasty hack called Portsupgrade.

    2. By Simon (217.157.132.75) on

      The OpenBSD developers do need to give the port tree at least some attention. A large number of ports need additional patches to run on OpenBSD, if these aren't made easily avaiable, I think many would choose a different operating system.

      There are nice things in the base system, but it isn't enough for many users, they need the port tree to provide easy access to third party software.

      Comments

      1. By Brad (67.70.23.103) brad at comstyle dot com on

        The ports tree does get attention and the patches are made easily available. They're in CVS with history. What more can you ask for?

        I don't see you pointing out any real problems. The point of your post was?

        Comments

        1. By Anonymous Coward (69.197.92.181) on

          Don't be so defensive. He's just telling the original poster that ports are still part of what makes openbsd, and just because somethings not in the base system, doesn't mean its not important. Without all the ports work, alot of people who currently use openbsd wouldn't.

          Comments

          1. By Simon (217.157.132.75) on

            Thank you

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]