KCLUG

kclug@kclug.org

3239 discussions

backing up symlinks
by Rob Becker 09 Nov '04

09 Nov '04

I'm playing with optimizing the startup scripts on an old redhat box. In order to avoid completely monkeying things up, I would like to back up the /etc/rc.d/rc3.d directory prior to making changes. I've read the man pages on both cp and ln and I'm coming up short regarding backing up the symlinks in this directory. Is there a way to create symlinks with the same names that point to the same files as the original symlinks? I know I can do it manually, but I'd prefer to use the * wildcard to get it all done in one shot if possible. Does anyone have any ideas? Am I thinking about this the wrong way? Please enlighten me if I am missing an obvious solution. Thanks. Rob

2 1

Re: Network cloning of raid array
by Brian Kelsay 09 Nov '04

09 Nov '04

Feels like I have this conversation every 6 months or so, but here goes. One of the benefits of using Ghost or partimage or g4u (ghost for Unix, forgot that one before), is that you are taking a snapshot of a *non-running* OS. Not so much of a problem on Linux as it is w/ Windows, but this gets rid of the problem of trying to capture locked files and log files tend to be in flux too. It lets you get an exact copy at a moment in time and copy over a network, to a partition or to another disk, possibly to an exact duplicate RAID system to create a redundant backup. Your idea of using a LiveCD is dead on, but the tarring and the order are probably off. Tar was originally meant to be used during backup to tape, so you may be OK using it, but I prefer using imaging tools that are able to lay down exact duplicate filesystems and compensate for larger or smaller destination partitions. They dynamically change the partition table to expand the image to fit the available space. Some programs do this well, Ghost, others do it poorly if at all, Partition Image by PowerQuest (not to be confused w/ partimage). Knoppix should have partimage and may have g4u. To get Ghost or Partition Image (recently bought out by Symantec, the owners of Ghost) you have to pay or have a good friend that uses it. ;-) Brian Kelsay >>> djgoku <> 11/09/04 11:25AM >>> Could you just tar up the system boot to a live cd mount / and untar? On Tue, 09 Nov 2004 11:21:03 -0600, Brad <brad(a)bradandkim.net> wrote: > On Tue, 2004-11-09 at 10:29, Brian Kelsay wrote: > > Not sure if it will work in a RAID setup, but the Linux world has partimage, to image partitions. I still haven't had a chance to test it, but it is supposed to work with NFS shares, partition to partition and disk to disk. There may be something in the RaidTools package that will do what you want. > > > > Cool, I will check that out and see if it helps...thanks. >

2 1

RE: Proposed Web site redesign (beta)
by Brian Kelsay 09 Nov '04

09 Nov '04

Damn, you beat me to it. On the punctuation thing. I didn't see where he used the word "critique" though. The first word of a sentence should be capitalized though. All kidding aside, you did a good job Tallen. Brian, XOOPS is a CMS (Content Management System), like PHPNuke, PostNuke, Slashdot, et al. I think it is written in PHP. Steven said he wanted to use that since he knew it and it is very flexible due to lots of modules being available. I am worried about speed issues, but the new LUG server seems to be very beefy, so far. I still prefer the listserv, because everybody has a mail client and it is usually open all day. Unfortunately, if I am only subscribed from home, then I may not be able to check those messages throughout the day unless I have webmail of some sort. We'll see how it shakes out. If someone else thinks they can do better, they can always volunteer and start w/ they're own space or collaborate w/ Steven. Brian Kelsay >>> "Brian Densmore" <> 11/09/04 09:42AM >>> > -----Original Message----- > From: Steven Hildreth > > ... Not to sound ungrateful for any comments/suggestions, but lets try to > make them constructive and productive to making a modern, updated, > technical site for our group. I know you Linux guys (myself > included) tend to argue over semantics versus content ;) Nice job. I posted a comment on the site. Not to be picky or nitpicky or argue over semantics, buuuuuuuut ... sentences usually end in one of [.?!] not ";)" . ^ Note period to end sentence. ________________| Also critique is spelled with a "u". ;') <snip> What is XOOPS? (Sure I could google it, but you brought it up - and I'm lazy today.)

4 3

WinPT
by djgoku 09 Nov '04

09 Nov '04

Anyone had much luck with WinPT? I am trying to get this thing working!!! I try to verify a file and just sits there. I created a gpg key though. Guess I could just copy it to my linux box at work. -- dj_goku -www.djgoku.com- -www.tektronic.org-

1 0

RE: Proposed Web site redesign (beta)
by Brian Densmore 09 Nov '04

09 Nov '04

> -----Original Message----- > From: Steven Hildreth > > ... Not to sound ungrateful for any comments/suggestions, but lets try to > make them constructive and productive to making a modern, updated, > technical site for our group. I know you you Linux guys (myself > included) tend to argue over semantics versus content ;) Nice job. I posted a comment on the site. Not to be picky or nitpicky or argue over semantics, buuuuuuuut ... sentences usually end in one of [.?!] not ";)" . ^ Note period to end sentence. ________________| Also critique is spelled with a "u". ;') All jesting aside. I like your design and of course it needs more content, which is irrelevant at this stage. I did notice two things that popped out at me. One, I like having a top pull down menu, not that the site needs one. Also, when going from the home page to other pages in the sidebar I noticed the links disappear (at least some of the time). It was quite noticeable and made me look off to the side momentarily taking my attention away from the content. That may be ok, seeing as you have a menu entry for links, but it may be more aesthetically pleasing to leave the links on all pages. What is XOOPS? (Sure I could google it, but you brought it up - and I'm lazy today.) $0.02, Brian

1 0

RE: chroot breakout (was: Xen 2.0 Virtual Machine)
by Garrett Goebel 09 Nov '04

09 Nov '04

Brian Densmore wrote: > >> Garrett Goebel wrote: >> >>> Have you tried to just chroot into another one? >> For _a_ test environment, that's fine. But not for running multiple >> simultaneous test environments, or giving away root accounts. > >Are you saying that you can't open up multiple CLIs and run chroot in >as many simultaneous instances as memory and diskspace allow? Sure, but instances of what? Processes not kernels. You couldn't for instance test the setup of a high availability cluster... >[somewhat OT: ] >Also if one can break out >of a chroot environment then they have the skill to own the machine >anyway. You need to be able to find and use a security flaw on the >machine that would give you root access and have access inside of >the chrooted environment to a compiler or perl interpreter. So the >fact that one could own a machine from inside a chroot environment >doesn't increase the possibility that someone could get root access. Unless of course you _want_ to give someone root access without fear that they can subvert their host. Chroot is fine for running services under a low privilege account in a jail. It isn't a cure-all. >Although what that has to do with wanting to run a VM, which is what >this thread is about, eludes me. In order to run a VM a user would >need an account on your box, and if they are going to crack your >system and have the knowledge to break out of a chrooted environment, >then they can own your box from their user account. In order to run a UML VM on a box, you need to run a UML instance which the end user could log into. They don't need _access_ to an account on the UML host. Except to the extent that UML instance would be running under some set of credentials. With UML I can give anyone I wish a root account on their own virtual Linux box... I still have to worry about them misusing it or being penetrated, but not so much about attempts to subvert the uml host. I think UML is promising choice for ISP's who offer co-hosting services. -- Garrett Goebel IS Development Specialist ScriptPro Direct: 913.403.5261 5828 Reeds Road Main: 913.384.1008 Mission, KS 66202 Fax: 913.384.2180 www.scriptpro.com garrett at scriptpro dot com

6 5

RE: chroot breakout (was: Xen 2.0 Virtual Machine)
by Brian Densmore 08 Nov '04

08 Nov '04

-----Original Message----- > From: Garrett Goebel > > Brian Densmore wrote: >> >>> Garrett Goebel wrote: >>> >>>> Have you tried to just chroot into another one? >>> For _a_ test environment, that's fine. But not for running multiple >>> simultaneous test environments, or giving away root accounts. >> >> Are you saying that you can't open up multiple CLIs and run chroot in >> as many simultaneous instances as memory and diskspace allow? > Sure, but instances of what? Processes not kernels. You couldn't for > instance test the setup of a high availability cluster... Well for high availability cluster emulation I would recommend using the King of VMs. IBM z390/Linux. ;') > Chroot is fine for running services under a low privilege account > in a jail. It isn't a cure-all. Very true. > With UML I can give anyone I wish a root account on their own virtual > Linux box... Sounds pretty good.

1 0

RE: Xen 2.0 Virtual Machine
by Garrett Goebel 08 Nov '04

08 Nov '04

Brian Kelsay wrote: > > I went and read most of the HOWTO on User-mode Linux and > I have to say, I'm not sure if it actually gives you a > good impression of a new kernel being able to run on your > hardware. The host kernel is still in the way. You use > virtual devices and filesystems. Sure, it's safe, but it > won't tell you definitively if the new kernel works on > your hardware. If you can boot to a UML kernel, you can work with whatever hardware the underlying kernel has compiled virtual device support for. The whole point is that UML doesn't work on hardware. It runs in usermode on top of a kernel which does. Actually, I'm lying ;) One of the directions UML is moving toward is the ability to compile support for selective direct hardware access. Certain hardware resources could be dedicated to different UML instances. This would allow you to decide where you want to place the virualization performance vs. security etc. tradeoffs. > Now if you want to test other software on a specific > kernel without risking crashing the system, then this > appears to be the answer you've been looking for. Or > if you want to test features of a new kernel, then this > would work. If you want to host an app like a webserver > in a controlled/limited environment, giving near root > privileges to a user, then this may work also, but a > chrooted jail might be better. Then again, I could be > wrong. Sure... use the right tool for the job. I just meant to underscore that chroot isn't always the right tool for the job. It depends on what you mean by a test environment, and whether or not the test environment needs root privileges. > User-mode appears to use less resources than a Xen, VMWare > or Bochs, as does chroot, but that is Linux on Linux only. > Each can have its place. You're preaching to the choir ;-) -- Garrett Goebel IS Development Specialist ScriptPro Direct: 913.403.5261 5828 Reeds Road Main: 913.384.1008 Mission, KS 66202 Fax: 913.384.2180 www.scriptpro.com garrett at scriptpro dot com

1 0

Re: Xen 2.0 Virtual Machine
by Brian Kelsay 08 Nov '04

08 Nov '04

I went and read most of the HOWTO on User-mode Linux and I have to say, I'm not sure if it actually gives you a good impression of a new kernel being able to run on your hardware. The host kernel is still in the way. You use virtual devices and filesystems. Sure, it's safe, but it won't tell you definitively if the new kernel works on your hardware. Now if you want to test other software on a specific kernel without risking crashing the system, then this appears to be the answer you've been looking for. Or if you want to test features of a new kernel, then this would work. If you want to host an app like a webserver in a controlled/limited environment, giving near root privileges to a user, then this may work also, but a chrooted jail might be better. Then again, I could be wrong. User-mode appears to use less resources than a Xen, VMWare or Bochs, as does chroot, but that is Linux on Linux only. Each can have its place. Brian Kelsay >>> "Brian Kelsay" <> 11/08/04 08:20AM >>> Are you possibly referring to User-Mode Linux? That is the only thing that comes close to the discussion at hand. http://user-mode-linux.sourceforge.net/ Not until I found that page did your post make sense. Brian Kelsay >>> David Nicol <davidnicol> 11/06/04 10:26PM >>> Isn't UML the preferred way to do that? On Fri, 05 Nov 2004 14:45:15 -0600, Thomas Bruno <crweb@> wrote: > It would be a better solution per cost, if all you wanted was to run > Linux on Linux.

1 0

RE: chroot breakout (was: Xen 2.0 Virtual Machine)
by Brian Densmore 08 Nov '04

08 Nov '04

-----Original Message----- > From: Garrett Goebel > >> Have you tried to just chroot into another one? > For _a_ test environment, that's fine. But not for > running multiple simultaneous test environments, > or giving away root accounts. Are you saying that you can't open up multiple CLIs and run chroot in as many simultaneous instances as memory and diskspace allow? [somewhat OT: ] Also if one can break out of a chroot environment then they have the skill to own the machine anyway. You need to be able to find and use a security flaw on the machine that would give you root access and have access inside of the chrooted environment to a compiler or perl interpreter. So the fact that one could own a machine from inside a chroot environment doesn't increase the possibility that someone could get root access. Although what that has to do with wanting to run a VM, which is what this thread is about, eludes me. In order to run a VM a user would need an account on your box, and if they are going to crack your system and have the knowledge to break out of a chrooted environment, then they can own your box from their user account.

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

KCLUG