Re: [gentoo-user] Re: X programs as root
Dale wrote: It appears that maybe portage got the order wrong or something was amiss with KDE. I re-emerged the following: kdebase-menu kdebase-menu-icons kmenuedit kdesu After that, it works again. So, either something was built in the wrong order or KDE had some issue with something. Oh, no config files to update after that either. Don't know what fixed it exactly but at least it works again. Maybe the OP should try rebuilding some packages too? May help, may not. Dale :-) :-) I spoke to soon I guess. I was gone for a while so I cut my puter off. Now, the programs don't work as root again even after recompiling those packages again. This is weird. Works one time then stops again. Dale :-) :-)
Re: [gentoo-user] Re: X programs as root
On Sunday 26 September 2010, Dale wrote: > Dale wrote: > > It appears that maybe portage got the order wrong or something was > > amiss with KDE. I re-emerged the following: > > > > kdebase-menu > > kdebase-menu-icons > > kmenuedit > > kdesu > > > > After that, it works again. So, either something was built in the > > wrong order or KDE had some issue with something. Oh, no config files > > to update after that either. > > > > Don't know what fixed it exactly but at least it works again. Maybe > > the OP should try rebuilding some packages too? May help, may not. > > > > Dale > > > > :-) :-) > > I spoke to soon I guess. I was gone for a while so I cut my puter off. > Now, the programs don't work as root again even after recompiling those > packages again. > > This is weird. Works one time then stops again. I always run several apps under kdesu and sometimes they don't start. Once I did an strace and found it hanging opening /var/tmp/kdecache-user; deleting all those solved the problem. HTH -Robin -- -- Robin Atwood. "Ship me somewheres east of Suez, where the best is like the worst, Where there ain't no Ten Commandments an' a man can raise a thirst" from "Mandalay" by Rudyard Kipling --
Re: [gentoo-user] Re: X programs as root
Robin Atwood wrote: On Sunday 26 September 2010, Dale wrote: Dale wrote: It appears that maybe portage got the order wrong or something was amiss with KDE. I re-emerged the following: kdebase-menu kdebase-menu-icons kmenuedit kdesu After that, it works again. So, either something was built in the wrong order or KDE had some issue with something. Oh, no config files to update after that either. Don't know what fixed it exactly but at least it works again. Maybe the OP should try rebuilding some packages too? May help, may not. Dale :-) :-) I spoke to soon I guess. I was gone for a while so I cut my puter off. Now, the programs don't work as root again even after recompiling those packages again. This is weird. Works one time then stops again. I always run several apps under kdesu and sometimes they don't start. Once I did an strace and found it hanging opening /var/tmp/kdecache-user; deleting all those solved the problem. HTH -Robin I was about to post this. I logged out of KDE and went to single user. I then used ps to list the processes and there was still a kdesu running. I killed it and a couple other left overs, then logged back in and everything is working again. It looks like KDE needs to do a little more clean up when logging out. I did clear some config files and cache files once before. I lost bookmarks somehow. I need to save those somewhere so I can restore them. ;-) Dale :-) :-)
Re: [gentoo-user] iwl5000 firmware fails to load
Running 2.6.35-r8. I suspected hal, as well. If I look through my /var/log/dmesg output, the load messages are almost identical to yours before I emerge the set. Unfortunately unmerging and re-emerging hal didn't seem to do anything. Then again -- would hal really have anything to do with the kernel loading the firmware? Hal shouldn't even be started by that point. -j On Sun, Sep 26, 2010 at 4:23 AM, Adam Carter wrote: > For reference here's mine > iwlagn: Intel(R) Wireless WiFi Link AGN driver for Linux, in-tree: > iwlagn: Copyright(c) 2003-2010 Intel Corporation > iwlagn :04:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 > iwlagn :04:00.0: setting latency timer to 64 > iwlagn :04:00.0: Detected Intel Wireless WiFi Link 5100AGN REV=0x54 > iwlagn :04:00.0: Tunable channels: 13 802.11bg, 24 802.11a channels > iwlagn :04:00.0: irq 32 for MSI/MSI-X > iwlagn :04:00.0: firmware: requesting iwlwifi-5000-2.ucode > iwlagn :04:00.0: loaded firmware version 8.24.2.12 > > Side note - If your system is pcie then you should turn on the MSI kernel > options. > > I have no idea really - tho out of the software you installed before the > breakage, Hal would have to be the biggest suspect. Maybe stop it form > starting at boot and reboot to see what happens?? (FWIW im running hald) I > assume you've added the hal use flag and done an emerge with newuse? > > What kernel are you running? > > >
[gentoo-user] Finalizing my backup system
I'm using backupninja to backup data from my laptop, desktop, and remote server onto a remote desktop system. backupninja is very simple and is really just an interface to a few other programs including rdiff-backup. I'm not worried about a good restore method for now, I want to focus on keeping it simple and protecting my data. This is the first time I've set up a real backup system and I'd love to get some advice from you guys. I've got a few questions. 1. This is the first time I've used passwordless SSH keys. root on each system being backed up logs into the remote desktop as a normal user to store the backups. Is this pretty safe? I suppose if root is compromised on any of the three systems being backed up (via physical access or otherwise), the remote desktop will also be compromised as a normal user. Maybe that normal user should be extraordinarily unprivileged? 2. backupninja can email reports. This works on my remote server which runs postfix, but my laptop and desktop don't run an MTA. Is there a simple one that would be well-suited to a purpose like this, or do I need full-blown postfix on my laptop and desktop? Whatever I choose, I'd also like to use it to send PORTAGE_ELOG messages from those systems. 3. On each system I back up /etc, /home/user/backup, /var/lib/portage/world, and /usr/src/linux/.config along with anything special from that system. Would anyone recommend I back up anything else? Some of the hidden directories in /home/user might come in handy, but I think I can rebuild those without too much trouble. 4. I have 600GB of music and photos that I'd like to back up somehow, but that is too much data to send to my remote desktop over my 20KB/s upload. How would you handle this? I was thinking maybe two external USB drives that I switch back and forth between being connected to the desktop system and being stored in a fireproof/waterproof container and hidden somewhere in my apartment to hopefully protect against fire, flood, and theft. 5. Do I have enough redundancy with backups only being stored on one system? 6. Any ideas for backing up the remote desktop which is the system where all the backups are stored? I can't back it up to my desktop or laptop because I'm behind some kind of a shared IP address. I also don't want to back it up to the remote server because that would require SSH keys on the remote server and if the remote desktop is compromised I don't want the remote server compromised along with it. Thanks guys, Grant
[gentoo-user] Re: Finalizing my backup system
On 09/26/2010 12:13 PM, Grant wrote: 2. backupninja can email reports. This works on my remote server which runs postfix, but my laptop and desktop don't run an MTA. Is there a simple one that would be well-suited to a purpose like this, or do I need full-blown postfix on my laptop and desktop? Whatever I choose, I'd also like to use it to send PORTAGE_ELOG messages from those systems. I use ssmtp for that purpose and there are at least two more similar packages in the mail-mta category, but I haven't tried them.
Re: [gentoo-user] Finalizing my backup system
On 09/26/2010 03:13 PM, Grant wrote: > I'm using backupninja to backup data from my laptop, desktop, and > remote server onto a remote desktop system. backupninja is very > simple and is really just an interface to a few other programs > including rdiff-backup. I'm not worried about a good restore method > for now, I want to focus on keeping it simple and protecting my data. > This is the first time I've set up a real backup system and I'd love > to get some advice from you guys. I've got a few questions. > > 1. This is the first time I've used passwordless SSH keys. root on > each system being backed up logs into the remote desktop as a normal > user to store the backups. Is this pretty safe? I suppose if root is > compromised on any of the three systems being backed up (via physical > access or otherwise), the remote desktop will also be compromised as a > normal user. Maybe that normal user should be extraordinarily > unprivileged? You can limit SSH access to only certain commands. On the remote desktop machine, you probably had to add an entry to the SSH authorized_keys file. You can prefix that line with the command that the user is allowed to run. For example (I use rdiff-backup too): command="/usr/bin/rdiff-backup --server",no-pty,no-port-forwarding ssh-rsa My uneducated guess is that this is safe unless there's a bug in rdiff-backup or ssh. > 2. backupninja can email reports. This works on my remote server > which runs postfix, but my laptop and desktop don't run an MTA. Is > there a simple one that would be well-suited to a purpose like this, > or do I need full-blown postfix on my laptop and desktop? Whatever I > choose, I'd also like to use it to send PORTAGE_ELOG messages from > those systems. Both ssmtp and nbsmtp work great for getting mail off your machine and to a real MTA. Just configure them with your account details (gmail or whatever). Create new account if you don't want to put your real details in a text file on every machine. > 3. On each system I back up /etc, /home/user/backup, > /var/lib/portage/world, and /usr/src/linux/.config along with anything > special from that system. Would anyone recommend I back up anything > else? Some of the hidden directories in /home/user might come in > handy, but I think I can rebuild those without too much trouble. Some programs stick stuff in /var/lib. The two most important that I know of are MySQL and PostgreSQL. I also back up all of /home and /root. > 4. I have 600GB of music and photos that I'd like to back up somehow, > but that is too much data to send to my remote desktop over my 20KB/s > upload. How would you handle this? I was thinking maybe two external > USB drives that I switch back and forth between being connected to the > desktop system and being stored in a fireproof/waterproof container > and hidden somewhere in my apartment to hopefully protect against > fire, flood, and theft. How often do you add new photos/music? You could make an initial trip with a big hard drive and copy everything that way. Then, the subsequent syncs would require a lot less traffic. > 5. Do I have enough redundancy with backups only being stored on one system? Depends on how important your stuff is. Ideally, you should rotate the physical media on your remote server and keep some copies off-site. That adds cost obviously; only you know whether or not it's worth it. > 6. Any ideas for backing up the remote desktop which is the system > where all the backups are stored? I can't back it up to my desktop or > laptop because I'm behind some kind of a shared IP address. I also > don't want to back it up to the remote server because that would > require SSH keys on the remote server and if the remote desktop is > compromised I don't want the remote server compromised along with it. If you really want to do it right, my recommendation would be to set up a separate machine with a swappable 2TB drive -- and buy a spare. Back up every other machine to it, and rotate the drive once a week or so. Keep the out-of-rotation drive at your house, office, or other safe place so a tyrannosaurus attack can't destroy all of the backups at once. If you locate the new machine close to your desktop/laptop, you can easily backup the photos and music to it too.
[gentoo-user] Boot Gentoo - by hand - line-by-line - command-by-command?
Hi, I'm curious if there's a way to capture the exact set of command that my machine executes as it boots up. I.e. - if I boot the machine, letting the boot loader find the kernel but then stopping at a bash prompt, before anything much has been done, I'd then like to know what set of commands I could use from that bash prompt to make the machine do whatever it does normally as it boots up. Is that possible, and is it documented anywhere? Or instead of capturing commands maybe just a list of things that happen and the bash commands I'd use to execute them myself. I think the init scripts, at least the ones I've found under runlevels, are more generic than I would like to run. I'm looking for something more like the portion of the install guide where we chroot into the new build, except executing that from the command line of a new machine that's booted. I.e.: mount /dev/hda4 /mnt/gentoo mount -t proc none /mnt/gentoo/proc mount -o bind /dev /mnt/gentoo/dev except modified to start from nothing and bring the machine up from scratch. Mount /proc, mount /dev, start udev, check root filesystem, etc. This is totally for self-education and nothing else really. Just curious about how it happens, what order things happen, etc. If anyone can recommend a good _basic_ book that talks about this I'd appreciate it. Thanks, Mark
Re: [gentoo-user] Boot Gentoo - by hand - line-by-line - command-by-command?
You may try Linux From Scratch and try to write the init script yourself, it's good for self-education Once you master the LFS you can learn more things interesting and amazing in gentoo On 9/27/10, Mark Knecht wrote: > Hi, >I'm curious if there's a way to capture the exact set of command > that my machine executes as it boots up. I.e. - if I boot the machine, > letting the boot loader find the kernel but then stopping at a bash > prompt, before anything much has been done, I'd then like to know what > set of commands I could use from that bash prompt to make the machine > do whatever it does normally as it boots up. > >Is that possible, and is it documented anywhere? Or instead of > capturing commands maybe just a list of things that happen and the > bash commands I'd use to execute them myself. > >I think the init scripts, at least the ones I've found under > runlevels, are more generic than I would like to run. I'm looking for > something more like the portion of the install guide where we chroot > into the new build, except executing that from the command line of a > new machine that's booted. I.e.: > > mount /dev/hda4 /mnt/gentoo > mount -t proc none /mnt/gentoo/proc > mount -o bind /dev /mnt/gentoo/dev > > except modified to start from nothing and bring the machine up from > scratch. Mount /proc, mount /dev, start udev, check root filesystem, > etc. > >This is totally for self-education and nothing else really. Just > curious about how it happens, what order things happen, etc. If anyone > can recommend a good _basic_ book that talks about this I'd appreciate > it. > > Thanks, > Mark > > -- @ghosTM55 Mechanism, not policy
Re: [gentoo-user] Boot Gentoo - by hand - line-by-line - command-by-command?
On Sun, Sep 26, 2010 at 5:43 PM, Thomas Yao wrote: > You may try Linux From Scratch and try to write the init script > yourself, it's good for self-education > > Once you master the LFS you can learn more things interesting and > amazing in gentoo > I was just looking at LFS in another tab. Thanks! > On 9/27/10, Mark Knecht wrote: >> Hi, >> I'm curious if there's a way to capture the exact set of command >> that my machine executes as it boots up. I.e. - if I boot the machine, >> letting the boot loader find the kernel but then stopping at a bash >> prompt, before anything much has been done, I'd then like to know what >> set of commands I could use from that bash prompt to make the machine >> do whatever it does normally as it boots up. >> >> Is that possible, and is it documented anywhere? Or instead of >> capturing commands maybe just a list of things that happen and the >> bash commands I'd use to execute them myself. >> >> I think the init scripts, at least the ones I've found under >> runlevels, are more generic than I would like to run. I'm looking for >> something more like the portion of the install guide where we chroot >> into the new build, except executing that from the command line of a >> new machine that's booted. I.e.: >> >> mount /dev/hda4 /mnt/gentoo >> mount -t proc none /mnt/gentoo/proc >> mount -o bind /dev /mnt/gentoo/dev >> >> except modified to start from nothing and bring the machine up from >> scratch. Mount /proc, mount /dev, start udev, check root filesystem, >> etc. >> >> This is totally for self-education and nothing else really. Just >> curious about how it happens, what order things happen, etc. If anyone >> can recommend a good _basic_ book that talks about this I'd appreciate >> it. >> >> Thanks, >> Mark >> >> > > > -- > @ghosTM55 > Mechanism, not policy > >
Re: [gentoo-user] Fire the fox.
On Sat, Sep 25, 2010 at 5:11 PM, Stroller wrote: > > On 25 Sep 2010, at 03:17, Kevin O'Gorman wrote: > >>> ... > > I've heard good things about it, but I'm under the impression it is > not free (as in beer). Is that true? > >>> I don't know but I can emerge -q icc > >> > >> There is other non-Free software you can install with Portage. > >> > >> Just yesterday I was looking at games-fps/ut2003 and games-fps/ut2004 - > I > >> believe these require the game's installer CDs to work. > >> > >> I would imagine that if you were to emerge ICC it would require an > >> activation key before it would compile anything, otherwise we'd all be > using > >> it. > >> > > > > Wouldn't that be kind of senseless since the source code is distributed? > > Knowing it would not be hard to bypass the activation key, if they wanted > > money for it they wouldn't let the source code out, license or no > license. > > Just because you can emerge a package doesn't mean the full source is > distributed. It could be a binary package, it could contain a small binary > blob for activation. > > Paul Hartman provides more info in his post of 24 September 2010 23:16:30 > GMT+01:00, but I was specifically replying to the assumption or implication > "if it can be emerged it must be free". > > You are right. Thanks for the clarification. ++ kevin -- Kevin O'Gorman, PhD
