Linked by Thom Holwerda on Mon 5th May 2008 17:12 UTC, submitted by Dale Smoker
OpenSolaris 2008.5, the new distribution based on the OpenSolaris operating system, has been released into the wild. This release follows the conventions set by many of the popular Linux distributions, such as being based on a single live CD with installer, but also adds a load of OpenSolaris-specific features such as ZFS, DTrace, Containers, and a new package management system, IPS. OpenSolaris 2008.5 is the fruit of Project Indiana.
Post a Comment
Member since:
2006-01-06
Member since:
2006-03-20
Doing a whole system update via web is pretty nice. pkg image-update creates a clone of your current ZFS root and updates that. Whether the update fails or you actually want to continue working as if nothing's happening, both is no problem. Reboot and the updated ZFS root is active.
This is called boot environments and can also be managed manually using beadm. Personally, I've set up the system as I expect it to be after a reinstall and customization, then cloned a BE from it. If something fubars, I can either boot that or better, create another clone from it and preserve the original.
From the management standpoint, it's pure joy.
Member since:
2006-09-01
And they're going to update the default shell to understand the backspace key
Thank god. Having the backspace key spew out useless junk instead of deleting characters is a personal irritant.
[I don't have any say in the process of shell selection, and I don't use Solaris enough to make arrangements to get around it.]
Member since:2005-07-07
Thank god. Having the backspace key spew out useless junk instead of deleting characters is a personal irritant.
I don't have any say in the process of shell selection, and I don't use Solaris enough to make arrangements to get around it.
I don't have any say in the process of shell selection, and I don't use Solaris enough to make arrangements to get around it.
It doesn't take too much time to type stty erase <back space> into your shell's .rc file.
Seriously people that have never used Unix suddenly think they are experts because they used linux. The basics is what Unix and Unix like Oses are all about.
Edited 2008-05-05 18:25 UTC
Member since:2005-07-24
Seriously people that have never used Unix suddenly think they are experts because they used linux.
Please ask yourself what constructive reason there was to say such a thing. It is perfectly reasonable to expect not to have to edit your profile to make the backspace key work in the year 2008. Instead of placing the blame on the user, fix the problem. I have been a Unix admin for 20 years (next month) and have watched the damage done by arrogant, RTFM, blame the user attitudes for that entire time. As such attitudes, and associated attitudes which spring from them, held Unix back from its rightful place on the desktop as another OS achieved greater and greater market success in that area. IMO, we can no longer afford to harbor such attitudes.
I'm glad to hear that OpenSolaris will be setting 'erase' properly by default, or whatever fix they are planning.
Trifles make perfection but perfection is no trifle.
Edited 2008-05-05 19:05 UTC
Member since:2005-08-27
Depending on who you ask, the function <backspace>, or the act of of deleting a single character to the left (previous), should be mapped the the rubout key, or <delete>.
If default Bourne shell was not mapped with "stty erase", which is an easy fix, its probably good cause behind this. There were many terminals in use so the path to least resistance was followed by allowing the user to set their own environment.
Member since:2005-07-06
Please ask yourself what constructive reason there was to say such a thing. It is perfectly reasonable to expect not to have to edit your profile to make the backspace key work in the year 2008.
Have to agree really. When you look at OpenSolaris, and its inability to do Virtual Terminals as well, you really have to ask what you're getting. All they're doing is stuff that Linux distros fixed......years ago.
It makes Linux look as ready for the desktop as it ever was, and it's actually a decent reminder just how much has improved.
Edited 2008-05-05 20:55 UTC
Member since:2005-07-07
Have to agree really. When you look at OpenSolaris, and its inability to do Virtual Terminals as well, you really have to ask what you're getting. All they're doing is stuff that Linux distros fixed......years ago.
Err let's see ZFS snapshotting before system updates seem like something no Linux distro supports.
It makes Linux look as ready for the desktop as it ever was, and it's actually a decent reminder just how much has improved.
Yet Linux has a minuscule desktop market share of 2.02%. MacOS X doesn't do Virtual Consoles either but has more market share, 7.3 %. MacOS X has been around since 2001. Linux since 1992. MacOS X
Bottom line virtual consoles don't sell anything.
Do you really have to pick at straws on every Sun related topic? It would make sense if you ever had a reasonable argument to make.
Member since:2005-08-26
When you look at OpenSolaris, and its inability to do Virtual Terminals as well, you really have to ask what you're getting.
It's not an inability, it's a lack of pre-configuration, just like with the backspace key. With that being said, give the guys time to iron out the bugs and add new features. They're already doing loads of work on the system, you can't expect it all in the first release. There has been a lot of work and effort put into getting this first release out.
Member since:2005-07-07
Please ask yourself what constructive reason there was to say such a thing.
The plain and simple reason of teaching users the fundamentals of using a system. Majority of the problems faced by computer systems users today would be solved with a little education.
You don't just give a kid,learning to ride for the the first time, a bike and tell him/her to go around a block for a spin, do you? You don't just hop on your bike and compete in the tour de france without training. How about when some one decides they want to drive? Similarly even if you know how to drive you just can't expect a person to drive around a race track.
If a person decides to use the shell on a Unix/linux systems they should at least know the basics of how to set shell properties.
It is perfectly reasonable to expect not to have to edit your profile to make the backspace key work in the year 2008.
Why? Why is it any more unreasonable than rm not moving a file to the trash can/recycle bin? The desktop environment has that as the mechanism for deleting a file why shouldn't the shell in 2008?
Instead of placing the blame on the user, fix the problem.
You are assuming it is a problem and not just a user preference.
I have been a Unix admin for 20 years (next month) and have watched the damage done by arrogant, RTFM, blame the user attitudes for that entire time. As such attitudes, and associated attitudes which spring from them, held Unix back from its rightful place on the desktop as another OS achieved greater and greater market success in that area. IMO, we can no longer afford to harbor such attitudes.
Unix had a lot more problems than user attitudes that prevented it from making it on the desktop. Linux has had a shell that does erase on backspace for almost a decade, how much desktop market share has it gained?
Member since:2005-07-06
The plain and simple reason of teaching users the fundamentals of using a system. Majority of the problems faced by computer systems users today would be solved with a little education.
If a system doesn't use the backspace to delete text then it is dead in the water. End of story. If you have to teach a user that that kind of 'fundamental' isn't what they think it is then you're in trouble.
Hey Solaris. Welcome to all the stuff Linux distros have learned over the past ten years (and how far they still have to go) ;-). It's an interesting ride.
You are assuming it is a problem and not just a user preference.
Oh dear.
Linux has had a shell that does erase on backspace for almost a decade, how much desktop market share has it gained?
More than Solaris, let's put it that way.
Member since:2008-05-06
"I have been a Unix admin for 20 years (next month) and have watched the damage done by arrogant, RTFM, blame the user attitudes for that entire time. As such attitudes, and associated attitudes which spring from them, held Unix back from its rightful place on the desktop as another OS achieved greater and greater market success in that area. IMO, we can no longer afford to harbor such attitudes.
Unix had a lot more problems than user attitudes that prevented it from making it on the desktop. Linux has had a shell that does erase on backspace for almost a decade, how much desktop market share has it gained? "
You completely missed the point. You said what you said to try and make yourself look cool/smart/whatever and make the other guy look dumb. I am pretty much new to GNU+Linux and Unix, so I can't say I've been able to personally witness this over the last 20 years, but let's just say it's hardly inconceivable that people like yourself have probably forced quite a few people away (from *nix) over the years (potential developers and users). It has nothing to do with having backspace delete characters to the left in the default terminal...
Member since:2006-12-28
Well said and I'm almost in total agreement, except on letting another OS take over the desktop ;-)
I personally find that a major issue with Unix, especially in the workstation market, was that none of them interacted very well with one another as well as the fact that they cost so much and mostly needed proprietary hardware. Sure, the BSD's where around but due to the AT&T court case, nobody wanted to touch em with a barge poll.
I must say that the amount of times I have heard users complain about the arrogant behavior of Unix admins is quite shocking.
Knowledge of Unix systems, especially shell scripting, is often worn like some kind of badge of honor and anyone not privy to the secret handshake must be some kind of dim wit. Failure to hide such dimwitedness, especially if the dimwit in question has the incredible stupidity to express a dislike, or even just ask a pertinent question, must be met with ridicule and denunciations as otherwise, how else will the admin be able to show off his incredible esoteric knowledge of Unix?
The arrogance is truly pathetic, not to mention unfounded, but unfortunately, I come across it far too often. It seems like you do too Steve.
Member since:2005-07-06
Knowledge of Unix systems, especially shell scripting, is often worn like some kind of badge of honor and anyone not privy to the secret handshake must be some kind of dim wit.
That's why Unix utterly and completely failed at its chance on the desktop around Windows 3.0 and OS/2. A bunch of people got together into a club, created a set 'standards' and assumed that because they had created a set of said 'standards' that it didn't matter what anyone else did, or how people expected their systems to work.
That attitude still remains today. Honestly, I ask you. 'Educating' users that the backspace button won't delete text?
Member since:2005-07-07
Knowledge of Unix systems, especially shell scripting, is often worn like some kind of badge of honor and anyone not privy to the secret handshake must be some kind of dim wit. Failure to hide such dimwitedness, especially if the dimwit in question has the incredible stupidity to express a dislike, or even just ask a pertinent question, must be met with ridicule and denunciations as otherwise, how else will the admin be able to show off his incredible esoteric knowledge of Unix?
Why would the dimwit be using the command line in the first place?
The pertinent question would be "How does one set the backspace to delete on the Bourne shell or default shell?".
Not "Oh come on don't tell me backspace doesn't work on this thing in 2008".
I find the later to not only be ignorant but extremely ignorant. As if someone is entitled to some thing.
The arrogance is truly pathetic, not to mention unfounded, but unfortunately, I come across it far too often. It seems like you do too Steve.
I am ok with such reactions from end users who have never used a similar system. But coming from linux users just rubs me the wrong way. That's my point.
Member since:2005-07-06
" The arrogance is truly pathetic, not to mention unfounded, but unfortunately, I come across it far too often. It seems like you do too Steve."
That is exactly the reason I do not use Linux in any serious way these days. The arrogance on the forums was intolerable. So if what you say is true it would appear that UNIX does not have a lock in that department.
I am downloading Open Solaris as I type and intend to give it a try. I may not like it, but just because you like the delete on backspace does not mean that every OS has to do that. It would be a rather dull world if every OS was the same. Anyway, if you want to use an OS you have to learn the quirks, and they all have them.
Member since:2005-08-15
Member since:2005-07-05
Member since:2006-09-01
It doesn't take too much time to type stty erase <back space> into your shell's .rc file.
If it would fail gracefully and not do anything that would be one thing, but it doesn't. It spits out characters.
Thanks for the info. I'll have to try that when I spin up a VM for Solaris. I do know that editing the .rc, or .cshrc, depending, file for the shell will load my preferences each time I login, but I haven't run into this problem outside of Solaris. Consequently, it's never been a priority for me to find a fix. Yes, I can read documentation and use a search engine.
Also, I couldn't implement the fix if I wanted to. I don't admin boxes, the standard for root is default, and there are 500+ boxes where I work. I don't want to change the shell when I'm work on recovering a crashed box. I'm not sure what will break, and I don't want to introduce variables.
There are perfectly good reasons I haven't found a solution to the problem, and none of them are related to me not knowing to the .rc file or my Linux usage.
Seriously people that have never used Unix suddenly think they are experts because they used linux. The basics is what Unix and Unix like Oses are all about.
I've never claimed to be a Unix expert, anywhere. I'm just happy Solaris will have an updated root shell, and I won't have to deal with that silliness if I decide to run it as my main OS.
I've used Unix, and Linux, Fedora, is the first Unix or Unix-like OS with the features to enable me to stick with it for a long period of time. I have to be able to get done what I need to get done. Which is the nice thing about Linux, it's Unix-like without making too many comfort sacrifices.
I use Linux to learn the fundamentals of Unix and Unix-like OSes. I threw threw myself into the deep end, so I couldn't retreat to a familiar OS. (I still have many OSes around, OS X, FreeBSD, Vista, XP, but I use Linux as my primary) Granted all the learning isn't one for one, but I learn the concepts.
Member since:2005-07-07
If it would fail gracefully and not do anything that would be one thing, but it doesn't. It spits out characters.
That's because the character is not mapped by the shell. It is not failing, per se. Just behaving differently.
Also, I couldn't implement the fix if I wanted to. I don't admin boxes, the standard for root is default, and there are 500+ boxes where I work. I don't want to change the shell when I'm work on recovering a crashed box. I'm not sure what will break, and I don't want to introduce variables.
You can exec bash at the prompt or tcsh or whatever shell you want.
There are perfectly good reasons I haven't found a solution to the problem, and none of them are related to me not knowing to the .rc file or my Linux usage.
That's beside the point.
I've never claimed to be a Unix expert, anywhere. I'm just happy Solaris will have an updated root shell, and I won't have to deal with that silliness if I decide to run it as my main OS.
If you are root on a box or 500+ at my company and called basic stuff like stty, silliness. I would be very worried.
I've used Unix, and Linux, Fedora, is the first Unix or Unix-like OS with the features to enable me to stick with it for a long period of time. I have to be able to get done what I need to get done. Which is the nice thing about Linux, it's Unix-like without making too many comfort sacrifices.
That's fine and totally understandable. But just because something is not what you are used doesn't mean it is broken. That was my point.
I use Linux to learn the fundamentals of Unix and Unix-like OSes. I threw threw myself into the deep end, so I couldn't retreat to a familiar OS. (I still have many OSes around, OS X, FreeBSD, Vista, XP, but I use Linux as my primary) Granted all the learning isn't one for one, but I learn the concepts.
Ah I remember the good old days when I was in college and made the same statement (running linux as my main Os and having others around to play with). :-) That's not a bad way to go about learning keep at it and all the best.
I must admit I was a little harsh and the discussion spun out of control.
Member since:2006-09-01
If you are root on a box or 500+ at my company and called basic stuff like stty, silliness. I would be very worried.
I'm still learning!
I work as Datacenter Site Support. If I'm root in a box, I'm taking commands from an Unix Engineer over the phone. I don't get free reign at all.
Thanks for pointing stty out. Now I know.
(Also, I was saying backspace not being mapped is silliness, not stty is silliness. Stty looks to be very useful. I wasn't clear on that one.)
That's fine and totally understandable. But just because something is not what you are used doesn't mean it is broken.
I understand it's personal preference, and my knowledge was the thing that was broken. You educated about the solution, and the problem, my knowledge, has been fixed.
I must admit I was a little harsh and the discussion spun out of control.
I my statements can be brash sometimes, so no problem. It was an interesting thread.
Member since:2005-07-24
If it would fail gracefully and not do anything that would be one thing, but it doesn't. It spits out characters.
It's certainly a different day and age today. :-) I've seen so much:
\E[23~
\[7m
\E[!p\E[?3;4l\E[4l\E>
in my time. Fixed so many broken terminfo and termcap definitions. Diagnosed so many flow control glitches and mismatched terminal settings on AT&T 4410s and 605s and Wyse50s and Wyse60s. Done so much blind typing into terinals whose screen contents were completely unrecognizable, that when I see a few strange characters, it takes my brain a moment to recognize that there is actually something wrong.
There are only really two important terminal types that I work with today, and those are "linux" and "xterm". And those work so consistently well these days that I can fully understand the surprise that the non-graybeards must experience when a key isn't mapped right. It's a sign that some of the more stupid crap I used to have to deal with is now truly history. Or at least a rarity. Praise the Lord! (And I'm an atheist!)
The backspace thing was, indeed, an embarrassing issue for the year 2008. Especially since, if I understand correctly, it was not a matter of the erase character not being set, but of the shell not handling the defined erase character properly. But it *does* seem to be history. I'm not sure what the old config was, but when I bring up gnome-terminal in 2008.05 I get bash, and the backspace works just fine. I was expecting to be able to ctrl-alt-F1 to check out a text console, but that doesn't seem to be the right key sequence. Unix is Unix is Unix[1]... except for all those little things we take for granted about our usual flavor. :0
[1] That would be "POSIX-like OS is POSIX-like OS is POSIX-like OS" for you anal retentives regarding Linux not being Unix. But even you have to admit it loses something in the translation. ;-)
Edited 2008-05-06 22:29 UTC
Member since:2005-07-07
There are only really two important terminal types that I work with today, and those are "linux" and "xterm". And those work so consistently well these days that I can fully understand the surprise that the non-graybeards must experience when a key isn't mapped right.
This almost as silly as an American going to london and complaining that the steering wheel is on the "wrong side" and they drive on the "wrong side" of the road and vice versa.
The backspace thing was, indeed, an embarrassing issue for the year 2008. Especially since, if I understand correctly, it was not a matter of the erase character not being set, but of the shell not handling the defined erase character properly.
ergo not understanding the hex code for backspace to mean erase. There is no properly here. If that particular shell revision always behaved that way by design then it isn't a bug. The user expectation is, in effect PEBKAC.
But it *does* seem to be history. I'm not sure what the old config was, but when I bring up gnome-terminal in 2008.05 I get bash, and the backspace works just fine.
Like I said mountian out of a mole hill.
I was expecting to be able to ctrl-alt-F1 to check out a text console, but that doesn't seem to be the right key sequence. Unix is Unix is Unix[1]... except for all those little things we take for granted about our usual flavor. :0
Which other major Unix (AIX, HP-UX) has a key sequence that switches to a virtual console?
That would be "POSIX-like OS is POSIX-like OS is POSIX-like OS" for you anal retentives regarding Linux not being Unix. But even you have to admit it loses something in the translation. ;-)
POSIX doesn't define the things you are complaining about.:-)
Edited 2008-05-06 22:42 UTC
Member since:2006-03-20
Thank god. Having the backspace key spew out useless junk instead of deleting characters is a personal irritant.
[I don't have any say in the process of shell selection, and I don't use Solaris enough to make arrangements to get around it.]
[I don't have any say in the process of shell selection, and I don't use Solaris enough to make arrangements to get around it.]
It's the old Bourne shell that doesn't support backspace. You're still free to use any other shell, you can change it with usermod.
Indiana comes with ksh93 replacing sh and as default shell for root, and bash as shell for regular users. Both do support backspace.
Member since:
2005-07-06
Member since:
2005-08-27
I installed nexenta on a Thinkpad, and though I liked the packaging system (from debian), I felt uneasy since it was unlikely that it would become the default or option for general release solaris.
IPS seems to be able to use Sun's or sunfreeware as package repositories. I hope this won't mean it is as limited. For all my solaris installs, I've been using pkgsrc (http://pkgsrc.org) from the NetBSD team, which has a much wider range of updated packages.
Member since:
2007-01-01
RE: OpenSolaris 2008.05 Screenshots
by kaiwai on Tue 6th May 2008 02:27 UTC
in reply to "OpenSolaris 2008.05 Screenshots"
Member since:2005-07-06
Thanks for the screeshots. I am confused, however, when I hear zfs being used, are they using ZFS for root/boot - whats the story with that? I've also avoided 2008.5 because it is based off b86 which includes a nasty Intel 4965 bug which results in crappy wireless performance.
RE[2]: OpenSolaris 2008.05 Screenshots
by Weeman on Tue 6th May 2008 10:00 UTC
in reply to "RE: OpenSolaris 2008.05 Screenshots"
Member since:2006-03-20
RE[3]: OpenSolaris 2008.05 Screenshots
by kaiwai on Tue 6th May 2008 10:56 UTC
in reply to "RE[2]: OpenSolaris 2008.05 Screenshots"
Member since:2005-07-06
"
Thanks for the screeshots. I am confused, however, when I hear zfs being used, are they using ZFS for root/boot - whats the story with that?
Thanks for the screeshots. I am confused, however, when I hear zfs being used, are they using ZFS for root/boot - whats the story with that?
ZFS root and boot and boot environments (clones). "
Is there an improvement in performance when compared to UFS? I've only ever used it briefly, but would sooner allow others to experience problems (and iron out those issues) before moving to it.
RE[4]: OpenSolaris 2008.05 Screenshots
by Weeman on Tue 6th May 2008 11:03 UTC
in reply to "RE[3]: OpenSolaris 2008.05 Screenshots"
Member since:2006-03-20
Member since:
2005-07-06
They really need a way to configure the bootloader setup during the install. There's no conceivable excuse in the world for having the installer nuke a pre-existing grub setup as the only behavior. None. Dunno if they thought streamlining it out would be a "good idea" or if someone got lazy on the technical side, but it's not even remotely amusing
Edited 2008-05-05 20:18 UTC
Member since:
2005-11-05
Who is OpenSolaris for?
I've had a quick look at the website and am none the wiser. There is no pitch to a market. The site seems to assume that the people who want to know all about OpenSolaris are the people who already know all about OpenSolaris.
This does not bode well. I've always liked the sound of OpenSolaris but why would I, or anyone else, want to use it?
Come on Sun, there must be at least one reason.
RE: Comment by moleskine
by wannabe geek on Mon 5th May 2008 23:53 UTC
in reply to "Comment by moleskine"
Member since:2006-09-27
RE: Comment by moleskine
by segedunum on Tue 6th May 2008 00:46 UTC
in reply to "Comment by moleskine"
Member since:2005-07-06
I've had a quick look at the website and am none the wiser.
You're not the only one.
Come on Sun, there must be at least one reason.
They want to attract Linux users, and especially contributors, who they expect to work for free on a project that they ultimately control to make the whole thing seem 'open source' so they can tell their customers that it's all just like Linux and there's no reason to panic (that's what it's mainly about):
http://mail.opensolaris.org/pipermail/ogb-discuss/2008-February/004...
This is hilarious:
http://mail.opensolaris.org/pipermail/ogb-discuss/2008-February/004...
They then proceed to have a go at Roy Fielding because he hasn't been enough of a member of the Solaris 'community'.
Basically, the whole thread is then one long argument between Sun employees and their own potential contributors(!) about the arrangement of the deckchairs on the Titanic and what Sun didn't promise when they started OpenSolaris.
Project Indiana now is OpenSolaris, and that's an internal Sun project with decisions made purely by Sun lead by Ian Murdoch where some source code might end up on the OpenSolaris site and servers and where you'll get a downloadable ISO. Don't get fooled, no matter how much they pretend otherwise.
RE[2]: Comment by moleskine
by Arun on Tue 6th May 2008 00:53 UTC
in reply to "RE: Comment by moleskine"
Member since:2005-07-07
Oh Looks OpenSolaris is just like linux then.
http://apcmag.com/why_i_quit_kernel_developer_con_kolivas.htm
RE[2]: Comment by moleskine
by Arun on Tue 6th May 2008 00:59 UTC
in reply to "RE: Comment by moleskine"
Member since:2005-07-07
Project Indiana now is OpenSolaris, and that's an internal Sun project with decisions made purely by Sun lead by Ian Murdoch where some source code might end up on the OpenSolaris site and servers and where you'll get a downloadable ISO. Don't get fooled, no matter how much they pretend otherwise.
Again your ineptitude in reading and comprehension fails you. The argument was about the OpenSolaris trademark for a distribution and nothing about code.
RE[3]: Comment by moleskine
by segedunum on Tue 6th May 2008 11:12 UTC
in reply to "RE[2]: Comment by moleskine"
Member since:2005-07-06
Again your ineptitude in reading and comprehension fails you.
Well, you can't even read what graphs are actually telling you when they actually have, you know, labels telling you exactly what they show and where what they show is actually described in the article:
http://www.osnews.com/permalink?312929
The argument was about the OpenSolaris trademark for a distribution and nothing about code.
I suggest you read the entire thread as I did (which is a waste of time where you're concerned, obviously). What Roy Fielding was talking about was an awful lot more than any trademark issues and promises (which Sun decided to go back on).
RE[4]: Comment by moleskine
by Arun on Tue 6th May 2008 15:42 UTC
in reply to "RE[3]: Comment by moleskine"
Member since:2005-07-07
[
Well, you can't even read what graphs are actually telling you when they actually have, you know, labels telling you exactly what they show and where what they show is actually described in the article:
http://www.osnews.com/permalink?312929
Well, you can't even read what graphs are actually telling you when they actually have, you know, labels telling you exactly what they show and where what they show is actually described in the article:
http://www.osnews.com/permalink?312929
Go back to that thread and read my response to your post.
I suggest you read the entire thread as I did (which is a waste of time where you're concerned, obviously). What Roy Fielding was talking about was an awful lot more than any trademark issues and promises (which Sun decided to go back on).
Go read that thread again. BTW I am done with your Anti-Sun bashing. Good riddance.
Member since:
2005-07-24
I'm waiting for my download to complete and figured I'd post something positive which might not turn into a nasty Linux vs Solaris brawl. (One can hope.) ;-)
Distributing this as a live CD was a good move. It has been a while since I have tried installing anything Solaris since, in the past, the install has gotten stuck early in the installation process. Hardware support is no doubt better than it once was, but if it were several CD's and then a regular install, I might not have decided to try it out. I suspect I am not alone. The live cd approach is definitely a good one for OpenSolaris.
RE: Live CD is a nice touch
by Arun on Tue 6th May 2008 01:26 UTC
in reply to "Live CD is a nice touch"
Member since:2005-07-07
I'm waiting for my download to complete and figured I'd post something positive which might not turn into a nasty Linux vs Solaris brawl. (One can hope.) ;-)
But wait hasn't ubuntu used this for a while. A modern OS in 2008 would be dead in the water if it didn't use a live CD. :-) Just kidding.
Give us an impression of how it goes. I haven't installed Solaris on an x86 machine since Solaris 7/8. I might just have to dust off my old pc to check it out. That's the problem with only having PPC macs at home.
RE[2]: Live CD is a nice touch
by sbergman27 on Tue 6th May 2008 01:42 UTC
in reply to "RE: Live CD is a nice touch"
Member since:2005-07-24
Give us an impression of how it goes. I haven't installed Solaris on an x86 machine since Solaris 7/8. I might just have to dust off my old pc to check it out. That's the problem with only having PPC macs at home.
Alas, it didn't get very far. Grub came up and looked quite nice. The little dots marched across the screen as it read in the kernel. I got a text screen with 3 lines ending in "Use is subject to license restrictions" or something like that, and then it froze. lspci output follows, in case someone spots an obvious problem. I do not have any sata or pata drives in this machine. (We had a multi-day power outage over the winter that convinced me of the advantages of a USB-based system.) Perhaps it's expecting to see an ata or scsi drive? Or maybe it doesn't like my nforce3 chipset? (Or maybe it's just the Steve Bergman curse on home Solaris installation.) This was a sort of spur of the moment decision, so I didn't check the hardware compatibility list before hand. (Hey, that's what live cd's are for!) ;-)
-----
00:00.0 Host bridge: nVidia Corporation nForce3 250Gb Host Bridge (rev a1)
00:01.0 ISA bridge: nVidia Corporation nForce3 250Gb LPC Bridge (rev a2)
00:01.1 SMBus: nVidia Corporation nForce 250Gb PCI System Management (rev a1)
00:02.0 USB Controller: nVidia Corporation CK8S USB Controller (rev a1)
00:02.1 USB Controller: nVidia Corporation CK8S USB Controller (rev a1)
00:02.2 USB Controller: nVidia Corporation nForce3 EHCI USB 2.0 Controller (rev a2)
00:06.0 Multimedia audio controller: nVidia Corporation nForce3 250Gb AC'97 Audio Controller (rev a1)
00:08.0 IDE interface: nVidia Corporation CK8S Parallel ATA Controller (v2.5) (rev a2)
00:0a.0 IDE interface: nVidia Corporation CK8S Serial ATA Controller (v2.5) (rev a2)
00:0b.0 PCI bridge: nVidia Corporation nForce3 250Gb AGP Host to PCI Bridge (rev a2)
00:0e.0 PCI bridge: nVidia Corporation nForce3 250Gb PCI-to-PCI Bridge (rev a2)
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] HyperTransport Technology Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscellaneous Control
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon R200 QM [Radeon 9100]
02:0b.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller (rev 13)
Edited 2008-05-06 01:44 UTC
RE[3]: Live CD is a nice touch
by Arun on Tue 6th May 2008 01:54 UTC
in reply to "RE[2]: Live CD is a nice touch"
Member since:2005-07-07
Alas, it didn't get very far. Grub came up and looked quite nice. The little dots marched across the screen as it read in the kernel. I got a text screen with 3 lines ending in "Use is subject to license restrictions" or something like that, and then it froze. lspci output follows, in case someone spots an obvious problem. I do not have any sata or pata drives in this machine. (We had a multi-day power outage over the winter that convinced me of the advantages of a USB-based system.) Perhaps it's expecting to see an ata or scsi drive? Or maybe it doesn't like my nforce3 chipset? (Or maybe it's just the Steve Bergman curse on home Solaris installation.) This was a sort of spur of the moment decision, so I didn't check the hardware compatibility list before hand. (Hey, that's what live cd's are for!) ;-)
Ouch that sucks!
Are you booting the text or the graphical version?
RE[4]: Live CD is a nice touch
by sbergman27 on Tue 6th May 2008 02:09 UTC
in reply to "RE[3]: Live CD is a nice touch"
Member since:2005-07-24
RE: Live CD is a nice touch
by shapeshifter on Tue 6th May 2008 02:28 UTC
in reply to "Live CD is a nice touch"
Member since:2006-09-19
RE[2]: Live CD is a nice touch
by oxygene on Tue 6th May 2008 07:33 UTC
in reply to "RE: Live CD is a nice touch"
Member since:2005-07-07
Member since:
2007-01-22
Member since:2007-03-30
Member since:
2005-07-06
So far so good. Their server let me have plenty of bandwidth and I got the download in about 10 minutes. The live CD booted up to a nice looking desktop. The live CD seemed to run slower than others I have used, but still worked just fine.
All in all it looks good enough to do an install and take a better look at it. It is a far cry from the last Solaris i looked at several years ago.
Member since:
2007-07-27
looks like a killer. You can do a snapshot, patch the system and when you boot - you can choose which snapshot to boot from GRUB. If the system behaves strange, just rollback in an instant. All snapshots are available via GRUB.
Imagine doing that in a production system. If something goes wrong, just rollback. If you get virus, just rollback. Version control, like CVS or SVN, for the entire system. Wow! Only ZFS allows this feature. And someone said that ZFS is just like another file system, albeit a little better? He seems to haven't understood anything. If he can't imagine the benefits of this, then his loss. Right?
Edited 2008-05-06 10:28 UTC
RE: Different Boot Environments
by segedunum on Tue 6th May 2008 11:05 UTC
in reply to "Different Boot Environments"
Member since:2005-07-06
Imagine doing that in a production system. If something goes wrong, just rollback.
Christ. This is why you have test environments (not that you can't do this on other systems), regardless of what you can and cannot snapshot on a production system...... I really worry about some people, and some of the people on the OpenSolaris lists are really worrying.
Yes, it's really lovely that ZFS is detecting some potential corruption that you didn't have when you had a working system before. It's also really lovely that ZFS is silently handling it. You still have to deal with this. The question is, why on Earth do you have that problem now where you didn't before, and why are people seeing the same thing with the same disk controller? Good luck with that one.
If you get virus, just rollback.
Roll back to what, exactly?! Do you know how far the virus went back? Have you just taken a snapshot of the infected system? How do you know it hasn't affected ZFS itself? Do you know what state the snapshot is in?
No. The only way you'll be able to make this work is through an adequate back up system (which snapshots simply cannot replace) with adequate checksumming on a per file and directory basis so you know exactly what has changed and when as a form of IDS.
Even then, it's possible to do much the same thing with LVM snapshots or Volume Shadow Copy.
Version control, like CVS or SVN, for the entire system. Wow! Only ZFS allows this feature.
Errr, no, because CVS, SVN or a backup system provide checksumming and diff changes on a per file and directory basis. Unless you make a snapshot for each and every change you don't have that with a snapshot system. If you do, you'll have to allocate a spectacular amount of space to account for divergences.
It's no replacement for a backup system.
RE[2]: Different Boot Environments
by Matt Giacomini on Tue 6th May 2008 17:38 UTC
in reply to "RE: Different Boot Environments"
Member since:2005-07-06
RE[2]: Different Boot Environments
by Arun on Tue 6th May 2008 18:18 UTC
in reply to "RE: Different Boot Environments"
Member since:2005-07-07
Christ. This is why you have test environments (not that you can't do this on other systems), regardless of what you can and cannot snapshot on a production system...... I really worry about some people, and some of the people on the OpenSolaris lists are really worrying.
Wow more FUD with no basic understanding.
Yes, it's really lovely that ZFS is detecting some potential corruption that you didn't have when you had a working system before. It's also really lovely that ZFS is silently handling it. You still have to deal with this. The question is, why on Earth do you have that problem now where you didn't before, and why are people seeing the same thing with the same disk controller? Good luck with that one.
Err. Because before they had nothing to detect the silent bit rot and now they do! Duh!.
Roll back to what, exactly?!
How about the first snapshot of a freshly installed system.
ZFS send can send a snapshot to a file and zfs recieve can restore it. You can have this snapshot on a completely different drive. Reinstall and restore from that snapshot. Much faster than setting everything up again. So you can send any number or snapshots to a different disk and maintain known good copies that are free from virii. If the snapshots in the system are not working out restore from the external disk. If you want this disk can be powered of when not in use.
Do you know how far the virus went back? Have you just taken a snapshot of the infected system? How do you know it hasn't affected ZFS itself? Do you know what state the snapshot is in?
None of that invalidates the benefits of ZFS snapshots and would hold true for even NetAPP filers. But the bottom line is such a system was not available for laptops or desktops in such a simple and extensive way.
No. The only way you'll be able to make this work is through an adequate back up system (which snapshots simply cannot replace) with adequate checksumming on a per file and directory basis so you know exactly what has changed and when as a form of IDS.
Even then, it's possible to do much the same thing with LVM snapshots or Volume Shadow Copy.
Even then, it's possible to do much the same thing with LVM snapshots or Volume Shadow Copy.
But ZFS makes that so easy. It tracks the changes when you snapshot, can compress the snapshots and even export it to a file. Let's see you do such a backup that with LVM and ext3. You'll need external tools and more work.
Volume shadow copy doesn't have the ability to export a snapshot or compress a snapshot.
Errr, no, because CVS, SVN or a backup system provide checksumming and diff changes on a per file and directory basis. Unless you make a snapshot for each and every change you don't have that with a snapshot system. If you do, you'll have to allocate a spectacular amount of space to account for divergences.
It's no replacement for a backup system.
It's not a replacement for a backup system but it enhances a backup system substantially. Your argument is it doesn't replace a back system so it is no good is pretty dubious and ignorant.
Edited 2008-05-06 18:19 UTC
RE[3]: Different Boot Environments
by segedunum on Tue 6th May 2008 22:44 UTC
in reply to "RE[2]: Different Boot Environments"
Member since:2005-07-06
"Christ. This is why you have test environments....
Wow more FUD with no basic understanding. "
So you don't have any test environments because ZFS is so great? Pffffffff......... That's what I thought, and what I thought when I wrote that comment, and that's the impression some of the OpenSolaris list topics have given me.
Lovely.
Err. Because before they had nothing to detect the silent bit rot and now they do! Duh!.
Errrrr, the problem is that there are people on that list, who are seemingly dumping working systems for ZFS, are encountering corruption issues that they never saw before with no indication of what on Earth caused them (did the move cause it?) or how to solve them now, and a lot of them are using the same drivers. Hmmmmmmm.
But hey, ZFS detected it, right?
ZFS send can send a snapshot to a file and zfs recieve can restore it. You can have this snapshot on a completely different drive. Reinstall and restore from that snapshot. Much faster than setting everything up again.
That's a bare metal image backup, right? This is new and cool because..........? I don't really know of anyone who reinstalls from scratch.
So you can send any number or snapshots to a different disk and maintain known good copies.....
Yep. Already doing it.
But the bottom line is such a system was not available for laptops or desktops