|
OpenBSD General Other questions regarding OpenBSD which do not fit in any of the categories below. |
|
Thread Tools | Display Modes |
|
|||
user suggestions to project
Quote:
It is due to the insular view taken adopted by its developers. It is a niche product developed by developers for themselves. I have a triple-boot system: OpenBSD, Debian and ArchLinux. OpenBSD 5.5 does not co-exist well with the other two. OpenBSD 5.6 fares slightly better, but not by much. The keyboard and mouse of my desktop machine do not function at times when you reboot from a Linux OS to OpenBSD. If you are rebooting from Debian or ArchLinux to OpenBSD on the same hard disk drive, you will encounter problems with OpenBSD's file system. Though OpenBSD, Debian and ArchLinux are classified as "Unix-like" OSes, OpenBSD's file system and its method of calculating and apportioning sectors and cylinders are unique. This might be due to the OS being a niche product. There is a noticeable degradation in performance of the hard disk drive if you dual-boot, triple-boot or quadruple-boot with OpenBSD and other distros. Agreed. This is due to the insular view of its developers. It is developed by developers for themselves. According to the Book Of PF, all of us end-users are just there for the ride. If you harbor so much as a dissenting view, you are persona non grata. I have a piece of advice for you: Do not ever dream of offering suggestions on how to improve the installer. I have done it before and been flamed. To the developers, the installer is super-perfect. You can't improve on perfection, can you? Last edited by cravuhaw2C; 2nd December 2014 at 09:02 PM. |
|
|||||
Quote:
Quote:
Quote:
AT&T's Unix and the Berkeley Software Distribution (BSD) which spawned from it both predate the IBM PC and its MBR partitioning schema. OpenBSD runs on 23 computing architectures today, but only 5 of those use MBRs. So it needs something other than MBR partitions for storage management, and it continues to use the disklabel system devised before MBRs came into existence. The fdisk(8) program used on MBR architectures allows the admin to use LBA addressing (sector numbers) or Cylinder/Head/Sector (CHS) addressing, because some older systems require the latter, they cannot use LBA addressing. There's nothing unique about CHS addressing for fixed-block-architecture disk drives, it's been around for many decades, even though for the last 20+ years, CHS values have been artificial. Quote:
If instead, you mean, "OpenBSD's disk performance sucks compared to any other OS," then at least I could guess that Soft Updates have not been enabled, and I would recommend enabling them (FAQ 14.6); they are a huge performance boost for the BSD Fast File System that OpenBSD uses. Soft Updates are not enabled by default, because OpenBSD has some architectures that cannot use them. Quote:
Last edited by jggimi; 2nd December 2014 at 10:49 PM. Reason: clarity. |
|
||||||
Ugh... I'm gonna respond to this. I'm not going to let FUD slide. And I'm certainly going to catch flak for this.
I meant what I said; I refuse to play semantic word games with you. Quote:
Quote:
Quote:
Quote:
Quote:
But you know what? Every once in a while someone sitting in the back decides they want to drive the bus. So they start, slowly, contributing to parts of OpenBSD. And in time, quality contributions earn that person the right to drive. And then guess what, that person has a say in the direction of OpenBSD. Every single developer starts from a "dissenting opinion" (OpenBSD was founded on having a dissenting opinion) and a drive to work to grow OpenBSD in a direction they see fit. If an anthropologist with no formal CS training (i.e., me) can become one of the drivers, then there is nothing stopping anyone else from becoming one too. Your problem, and this is a constant issue I see with people like you bent on spreading FUD, is that you are too willing to see structures and organizations as static entities. They're not: they're awesomely dynamic structures that anyone who wants to can get involved with and have amazing experiences--whether they want to be "just a user" or a developer or anything in between. It's the FUD that's going to stop those people from even starting. So sorry, but I am not going to let you do that. Quote:
If someone has quality, constructive improvements to any part of the system, then they should not hesitate to post such an improvement to one of the lists. Patches, of course, are better than just suggestions. |
|
||||
Quote:
I have been UNIX user since my high-school years (circa 1990) and one of the lucky few who made a jump from Atari to MicroVAX 3100 (Tru64 UNIX) consequently never having to deal with Wintel crap. Following industry trends I eventually gave up on Sparc64 and SGI and moved on Intel hardware 7-8 years ago. About the same time my UNIX usage pattern changed. UNIX became my hobby not just a tool I used for my daily work. Eventually hobby became a paid job. As somebody whose job description includes installation of UNIX and UNIX-like systems I feel I can say something about installers. A word "good" when it comes to OS installers have completely different meaning for professionals and casual user (even hobbyist). A good installer for a paid professional like myself have to satisfy many things. I will try to list some of my requirements in the order they come to my mind fully aware that I will miss few. Good OS installer has to: 1. Give me as much control as possible over the hardware. 2. Give me as much control over the disk sectors as possible 3. Allow me to install OS on RAID1 for redundancy. 4. Allow me to encrypt root partition. 5. Allow me to create software RAID or ZFS volumes. 6. Allow me to configure network, DNS and time synchronization daemon. 7. Allow me to quickly create administrative account and enable ssh services. 8. It should be scriptable so that I can quickly customize installation 9. Allow me automatically to perform unattended operating system installation and configuration including but not limited to turning on LDAP and mounting NFS shares. 10. It has to be fast because I have only 8h to install those 200 computing nodes. 11. It has to be extremely well documented. 12. Finally installer should be non destructive (no splash screen bullshit preferably no GUI) Now how many people at home do have to install OS let alone preform 200 installations a day (I actually work in very heterogeneous custom environment and have never done really more than 4-5 computers in a single day)? So lets see good and bad of various installers. I will start with Red Hat 6 release. Kickstart stellar. Real installer not so much. Apart of the fact that is a resources hog creating custom partition using GUI and mandatory mouse is royal pan in the ass. I have trouble booting on more than one occasion from a boot block which was installed on RAID 1 as oppose to a single member of the RAID 1. Encrypting large drives with LUKS takes ages. Finally you have to do after boot configuration (enter user etc). Is installer usable. Sure in particular Kickstart. Red Hat 7.0 branch have made creating custom partitions even more difficult. I would refer to new GUI is regression. It doesn't add anything really to people who have not used to old installer but it is a big distraction to somebody who used the tool. An example of really good installer is DragonFly BSD. Really stellar peace of the work. Creating custom partitions is breeze. Encryption (DragonFly TrueCrypt) is built in into installer. I have not done automated unattended installation though. It even allows you to test hardware in live mode or use it to repair existing partition. I should say that Red Hat allows you to do some repairs on damaged system but leaves much to be desired. TrueOS (PCBSD no GUI installer). Great for installing OS on ZFS mirror. Actually great for creating ZFS data storages. Never tried unattended installation and scripting. I am not sure it is even possible. It is really good for novice users not so great for mass deployment. FreeBSD old ncurses installer not my cup of tea. The new ncourses installer 10.xxx much worse. It requires a mouse for GUI Flipping through pages and pages of options click back click forth. Default UFS partition unacceptable. Installation on ZFS mirror "experimental". No way to encrypt entire thing during the installation (to my knowledge). OpenBSD installer comparing to all of above is my favourite besides DragonFly BSD installer. The only non-trivial thing which require you to drop into shell is installing on RAID 1 or even better installing on full encrypted RAID 1. Automated installation got really good over few past releases and I am not talking here about site.tgz. I am talking about Kickstart like thing. I installed NetBSD so long ago that I barely remember installer. I recall that default partitions Code:
/ /swap /home FreeNAS and PfSenense are similar to install. They are very involving and I am not aware that automatic unattended installation is possible. New FreeNAS will finally support installation on ZFS which is great. However that initial configuration script annoys the bones out of me. I have installed many Linux distros mostly for friends rather than work related. IIRC Slackware required third party tool for creating partitions before using actual installer. I learned a lot but it was royal pain in the ass. Ubuntu was typical click click guided installer. Great for a home user nightmare for a system admin. Actually the best Ubuntu installer is vulnerable Code:
dd Last edited by Oko; 3rd December 2014 at 08:12 PM. |
|
||||||||
Since you regard what I wrote in post #12 as FUD, please stop reading below this line. Do NOT take it the wrong way: I am helping you save time.
Quote:
Quote:
If you manage to read this far, brace yourself for a big shock: I do NOT know how to file bug reports. Quote:
Android OS, IOS and Microsoft Windows OS are examples of non-niche products. You could see there have been lots of creative/innovative energies poured into them. Quote:
Quote:
OK, OK...I get it. I know my place in this pecking order. If you say my post #12 contains FUD, please suppress (read: delete) it. I won't lose my sleep over it. Quote:
Quote:
Quote:
Oh my bad, please disregard it. I forgot you do NOT like to be drawn into a game of semantics. (sarcasm on the word "game") |
|
|||
ibara has labelled my post as FUD.
As such I see no point in providing more feedback as the label MUD might be attached to posts that I write in the future. |
|
|||
Hi guys,
I see NO point in taking part in discussions on this thread. As such I won't be responding to your posts. As I have done before and I am doing now, I shall take a long break from this forum. |
|
|||
Oko,
I have difficulty in trying to understand what you wrote. Don't take it the wrong way: either you improve your written English or write in your native language. |
|
|||||||||
Quote:
Quote:
Right there in the FAQ. Easily visible from the FAQ main page. Quote:
Quote:
Quote:
Quote:
Quote:
Quote:
Quote:
What is relevant is the project's goals. One of those goals is to allow installation of OpenBSD from floppy disks. This explanation is but one part of a complex matrix of design and philosophical decisions that have resulted in the OpenBSD installer looking the way it does. Sure! This is an example of a non-constructive attempt: "Your installer sucks because it doesn't have a GUI. The only way I could take OpenBSD seriously is if it had a GUI installer." This is an example of a constructive attempt: "OpenBSD lacks a GUI installer. I tried reading the mailing lists but have found no explanation as to why. I would like to work on such an installer, are there any caveats to doing this?" (Ignore the fact that in this case, there are discussions on the mailing lists as to why there is no GUI installer.) This constructive attempt doesn't even have code in it! This would be a nearly ideal attempt: "The installer is confusing on this step, here's some wording changes that I think would make it better." And if you could include a diff including that wording, it would be ideal. My objection was based on your intentional conflation of synonyms. |
|
||||
Quote:
(Hopefully you read this whenever you return.) * Not trying to insult Arch, but it is essentially a hobbyist system for people who like to tinker and have the latest shiny new sh*t. That neither makes Arch good nor bad, but it is definitely for a niche market. Last edited by sacerdos_daemonis; 3rd December 2014 at 05:21 AM. |
|
|||
This discussion has been split from its parent thread:
http://daemonforums.org/showthread.p...2394#post52394 ...as the subject has digressed from multi-booting with other Unix-like operating systems to a perceived unresponsiveness of the OpenBSD project to suggestions from the userbase. The innate value of this thread had nearly been exhausted, & it is reaching the point of lapsing into name-calling. While in my personal opinion this thread warrants locking so all involved can cool down, I also don't want to employ heavy-handed tactics to squelch further discussion on a public forum. However, I ask those composing new responses to maintain a constructive air. |
|
|||
Quote:
Being a niche product -whether true or false- is meaningless to my daily reality as a non-nerd user who needs a clean,secure,free,open-source,well-documented OS. Last edited by ocicat; 3rd December 2014 at 04:55 PM. Reason: Please use [quote] & [/quote] tags. |
|
||||
This complaint is a typical of the hordes of people that believe an operating system is only good if it can be installed and used by point-and-clicking with a pretty GUI, and a system that requires a little thinking and a modicum of reading are only suitable for programmers. Linux fora are inundated with such complaints. Fortunately, such complaints appear to be rare here. An appropriate response may be to refer those people to PCBSD. Easy installation, they have their own forum and it seems to be targeted at "beginners".
|
|
|||
An example of a constructive approach to improve an installer:
While ZFS has existed for years on FreeBSD, it was not possible to do a ZFS FreeBSD install until recently.RE: graphical installer The problem with a graphical installer or any GUI program is that it cannot be scripted or automated easily. Since a couple of weeks I have playing with Virtual Machine Manager, that is a GUI for creating and managing virtualized guests on Linux. After a while you get fed up because you are afraid to develop RSI from having to use your mouse so much. So I spend a week studying and reading the man pages of the command line utilities that allow you to do the same thing and even more than the Virtual Machine Manager GUI. I now can install a new OpenBSD virtual machine within 5 or 10 minutes. BTW in the 1980's I installed Forth on an Apple II computer by typing in a IIRC 50 page long 6502 assembler listing. It cost me more than a week to get that assembled and running
__________________
You don't need to be a genius to debug a pf.conf firewall ruleset, you just need the guts to run tcpdump |
|
||||
Quote:
Quote:
The difference between OpenBSD and some other projects in this respect is that OpenBSD is upfront about it. Or are you seriously suggesting that as an "end user" you have a say in what happens in the Debian GNU/Linux project or Arch Linux? Debian is a do-ocracy, if you don't write code, donate money or equipment/resources you don't get a say. Same with Arch Linux. The Linux Foundation itself, which employs Torvalds and Greg KH is funded by corporate backers, it is not some kind of charity producing free software for the masses out of love for mankind. You, I nor anyone else has commit rights to the Linux kernel - "you're along for the ride". There are graphical installers, because someone saw a need, sat down and wrote the code or was paid to do so. Not because many end users procrastinated incessantly until they were written for free. |
|
||||
Quote:
__________________
Linux/Network-Security Engineer by Profession. OpenBSD user by choice. |
|
||||
"It is due to the insular view taken adopted by its developers."
Of all things cravuhaw2c has said here, I think an "insular view taken by developers", is somewhat fair. Defined: Code:
in·su·lar adjective \ˈin(t)-su̇-lər, -syu̇-, ˈin-shə-lər\ : separated from other people or cultures : not knowing or interested in new or different ideas Full Definition of INSULAR 1 a : of, relating to, or constituting an island b : dwelling or situated on an island <insular residents> 2 : characteristic of an isolated people; especially : being, having, or reflecting a narrow provincial viewpoint OpenBSD, does however require active literacy at a level I recently thought to be more common. Consider this search for "42% of college graduates never read another book after college" https://duckduckgo.com/?q=42%25+of+c...+after+college Read the faqs, read the man pages, read the lists, buy the book (if able), use a text based installer - these things are unfortunately, beyond the comfort level of many people and that is not a criticism of OpenBSD. I went to the mail list with a problem and I had the lucky distinction of being flamed by one fellow and subsequently defended by Theo himself and it was, if anything, at best, a nuanced fault in a man page. There were however three things I didn't do: 1. Blame the software and community. 2. Fail to make a reasonable effort. 3. Presume to know something I didn't or make assertions for the community. Personally, I wonder if a dual boot Faq, simply exceeds it's scope. Archlinux, is a tinkerer's distro, linux from scratch (should do well), slackware, and what else don't provide much for the typical "Redhat, fedora, ubuntu, debian" dual boot crowd. I assume dual booting the BSDs are rather straight forward as well. The point being that a faq would cater to a niche user of OpenBSD. just my $.02, but I think cravuhaw2c isn't guilty of the first two and those bug reports are worth posting. Last edited by vanGrimoire; 13th December 2014 at 05:33 AM. Reason: code tags, broken |
|
|||
Quote:
He's definitely "guilty" of 3 and likely "guilty" of 1. What remains unclear is if he was blaming both software and community or just one facet. (or just me personally, which is fine too) |
|
|||
I am locking this thread, as a number of members are posting to it weeks after the initial (heated) discussion -- usually taking conversation in directions not central to the original poster's point(s).
Members who are compelled to comment on issues brought up in this thread are encouraged to start new threads. |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Need suggestions on what to name this project | TerryP | Off-Topic | 10 | 6th November 2010 03:13 PM |
Suggestions for my honours degree project... | scotsman | FreeBSD General | 7 | 20th September 2008 01:38 PM |
Suggestions for Web Traffic Logging? | Bruco | FreeBSD Ports and Packages | 16 | 18th September 2008 10:54 PM |
Mini-ITX motherboard suggestions | twisted_steel | General Hardware | 28 | 18th August 2008 09:32 PM |
Software suggestions | rex | FreeBSD General | 10 | 17th May 2008 12:00 AM |