Re: [gentoo-user] Time to move on?
Mike Diehl wrote: On Saturday 21 March 2009 21:00:11 Dale wrote: Mike Kazantsev wrote: On Sat, 21 Mar 2009 19:17:53 -0600 Mike Diehl mdi...@diehlnet.com wrote: Has Gentoo become such a moving target that it's no longer suitable for normal, every day, usage? If you're prepared to update you system at least once a week and have up-to-date knowledge of all the installed stuff, so you can at least make a decision whether you need some functionality or not... Then yep, I'd suggest gentoo. If you don't care about either then I don't understand why you started using it in first place - red hat or debian-based distro would've been much easier and simplier. I don't know if this is still the case or not but Mandrake updates seemed like a reinstall on top of itself to me. Sort of like when you reinstall windoze. It doesn't delete anything, user wise anyway, but just puts all the new stuff in there. You don't get the latest updates with Mandrake like Gentoo does but that doesn't appear to be to important to you since you don't update very often anyway. I suspect some other distro may better suite your needs. I been using Gentoo for years and update at least weekly and I rarely have trouble. However, if you let the updates pile up, you can have issues that are difficult to deal with. Overall, I agree with Mike here. Update regularly or use some other distro as he mentioned. Dale :-) :-) Ok, when I started using Gentoo, I remember a discussion about how often to do an emege world and the prevailing wisdom at the time was to do it when you needed a new feature, or fix. If the new wisdom is to update, say, weekly, I can live with that on the local machines here at the home/office. I'm a bit concerned about the servers I have co-located out of state, though. On the other hand, those are production machines and probably don't need to be upgraded many times during their lifetime. I've run several other distributions over the years and up until recently I've never looked back from Gentoo. I ran Slackware back when it came on 3.5 floppies. Of course it had NO package manager, so when Redhat hit the scene, I converted. Redhat, back then was built for a generic 486, so when Mandrake came along with pentium optimizations, I converted. But like you said, upgrading Redhat/Mandrake always seemed a bit windoze'ish to me. You really were simply piling the upgrade on top of the old system, like you said earlier. I used Suse on a project at work and hated every minute of it, and the help forums were mostly flamefests. Never even considered Suse for real work. Like I said, I've been using Gentoo for years now. When I met Daniel Robbins, I'd already been using Gentoo for several months. Gentoo is still the most customizable and optimize-able distribution available. Sometimes it's down right elegant. http://www.linuxjournal.com/article/10106 However, lately, Gentoo seems to have been plagued with problems. Circular blockers. 32/64 bit libraries. Package re-organization. Others. So here is the question: Are these just growing pains, or is this the trend with Gentoo? If I resolve to update frequently, will these problems become more rare? I'll start a new thread to seek help with my MythTV upgrade problem. Thanks for listening. Mike. This is my opinion and I am not a dev by any means. I think Gentoo is having some growing pains. I also think it is making huge leaps right now and they are really making some serious improvements. The newer portage will handle most blocks without you doing anything. There may be some exceptions to that but I would say the vast majority of blocks will be dealt with automagically. They seam to have came up with a way for portage to handle those blocks that is pretty seamless. That said, reading the elog or the messages after a emerge could be more critical. I read where someone may have missed a message and rebooted only to find that something was screwy and would no longer boot. I'm not sure they were running stable but either way these things can crop up. From what they posted, they had to boot with the CD and fix it. I sort of like that part about Gentoo. So, while portage may handle a lot for you, you need to run etc-update or whatever you use to update configs after each update or before you reboot at least. I run a single desktop machine here that runs folding and is my surfing machine. I could probably go a couple weeks between updates perhaps even a month and be OK. I think one to two weeks just seems to be a sweet spot for me at least. Long enough that you are not constantly updating but often enough that you are up to date. That would be especially true with regards to Mandrake, or whatever it is called now, and some others that take a while to update. They may be doing more testing or something but takes longer still.
[gentoo-user] Re: extending /usr partition...
On Sunday 22 March 2009, Albert Hopkins wrote: On Sat, 2009-03-21 at 14:13 -0700, BRM wrote: With all the words of LVM2 going on, I feel it is only appropriate to also mention the risk. On a desktop I had installed LVM2 considering that I did need to upgrade partitions every now and then and my previous solution was add another drive/partition and cross mount - e.g. like done with /usr/local under /usr, which worked fairly well. LVM2 worked great - until one of the drives crashed and I was trying to figure out what was on it. From that pov, volume management is a pain. I did figure out what I had mounted to it - but only after deconstructing the LVM configuration file to match it up with what I had put there. (And no, I had not yet gotten to doing an LVM soft-RAID solution to map a single LVM partition to two drives, which would certainly have helped.) I got my system working by adding a new drive that was not part of the volume group, and removing the old drives from the volume group. Fortunately, I had my volume setup so that they one partition was not made up of non-overlaping partitions on different drives. (e.g. partition A = sda1 + sda2 instead of sda1 + sdb1.) So, unless you are looking to use LVM in a soft-RAID solution between multiple physical drives, not multiple partitions on the same drive, (e.g. partition A = sda1 + sda2, with mirror on sdb1+sdb2), then I would not suggest it as should anything happen, it'll make data recovery that much harder. Just 2 cents for the pot. With or without LVM if you lose a drive then you've lost the data on it. LVM does have the capability of assembling a partially damaged volume group just not a partially damaged logical volume which, when you think about it, makes sense. And you can also throw in the standard warning about backing up your data. The point is that LVM adds an extra layer of complexity. I used LVM paired with soft RAID, and when I needed to boot from a liveCD I discovered that I had to rebuild the setup by hand. When you're in trouble it is pristine to have a quick way out instead of being swamped. I had my notes and managed to reckon the configuration (cold sweating!), but at the first occasion I reverted my system to plain RAID. Never used LVM for the few Gentoo server I manage. That said backup+RAID is the way to go. Cheers Francesco -- Linux Version 2.6.28-gentoo-r3, Compiled #1 SMP PREEMPT Sun Mar 8 12:38:59 CET 2009 Two 1GHz AMD Athlon 64 Processors, 4GB RAM, 4018.04 Bogomips Total aemaeth
[gentoo-user] Re: extending /usr partition...
On Saturday 21 March 2009, Alan McKinnon wrote: It's correct, and it also highlights just what a PITA it is to manipulate traditional disk partitions. With lvm, this becomes a breeze. With ZFS (we might see it one day) this becomes invisible. I thought it was already there: sys-fs/zfs-fuse Ciao Francesco -- Linux Version 2.6.28-gentoo-r3, Compiled #1 SMP PREEMPT Sun Mar 8 12:38:59 CET 2009 Two 1GHz AMD Athlon 64 Processors, 4GB RAM, 4018.04 Bogomips Total aemaeth
Re: [gentoo-user] Time to move on?
090321 Mike Diehl wrote:
Ok now, I'm getting fed up with all of the breakage
that I've seen in Gentoo in the last few months.
I haven't experienced any such thing.
I'm trying to upgrade MythTV.
I don't use that, so can't help directly.
Emerge told me to upgrade my profile, which I did.
I have (for a 64-bit system with an Intel Core 2 Duo processor) :
make.profile - ..//usr/portage/profiles/default/linux/amd64/2008.0
Now I'm doing an emerge -u world.
I never do that: I always do 'emerge -Dup world',
then decide which packages to update emerge them individually.
I also have a list of all the pkgs I have installed with dates + deps,
which I keep upto-date by hand as I emerge items.
I've never understood why 'emerge world' is considered standard:
repeatedly, there are appeals for help here resulting from its shortcomings
(was it copied from Free BSD when Gentoo was originally created ? ).
But before I could do that, I had to upgrade portage, with made sense.
When I go to emerge -u portage, I'm told:
sys-apps/mktemp (is blocking sys-apps/coreutils-7.1)
My hand-made list of pkgs tells me that I removed Mktemp 080419
that it had been required for Debianutils,
of which my current version is 2.28.5 installed 090314 .
Others had problems with this block, perhaps 1 year ago,
so if you really are that far behind in updating,
you should search the list archive to see what the advice was back then:
IIRC a new version of Debianutils incorporated the Mktemp stuff,
so they became incompatible.
So I do 'emerge -C mktemp' got a whole page of error messages.
The most basic msgs indicates the system can't load libselinux.so.1.
Do you have an item in 'make.conf' which requires that somewhere ?
Have you run Revdep-rebuild (pretend), to see what needs updating ?
'slocate' finds no similar file on my system.
All I want to do is upgrade a machine that I built a few months ago.
A few months can be a long time in the Gentoo world (smile).
For a desktop machine, you should do a full update = once/month :
I do it as a matter of routine every Saturday ('eix-sync' + follow-up).
As others have advised, Gentoo is not for people
who want to install forget: for that, try Mandriva, a respectable distro.
Anyway, I've offered a few hints above: try them ask again here.
--
,,
SUPPORT ___//___, Philip Webb
ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto
TRANSIT`-O--O---' purslowatchassdotutorontodotca
[gentoo-user] Re: Time to move on?
On Sunday 22 March 2009, Mike Diehl wrote: Now I'm doing an emerge -u world. But before I could do that, I had to upgrade portage, with made sense. When I go to emerge -u portage, I'm told: sys-apps/mktemp (is blocking sys-apps/coreutils-7.1) So I do: emerge -C mktemp Now I've gotten a whole page of error messages. The most basic of error messages indicates that the system can't load libselinux.so.1. I'm not using SElinux Nor do I want to. I've read the other threads; are you still interested in finding the solution to this showstopper ? My guess is that the system was already unstable, maybe awaiting for a revdep-rebuild. Are you completely blocked or it's possible to fix at last the basic elements? Ciao Francesco -- Linux Version 2.6.28-gentoo-r3, Compiled #1 SMP PREEMPT Sun Mar 8 12:38:59 CET 2009 Two 1GHz AMD Athlon 64 Processors, 4GB RAM, 4018.04 Bogomips Total aemaeth
[gentoo-user] [OT] splitting and printing big image
Hi! I have a PDF document. One of documet's page is a (scalable) schematics. Printing the page via my A4 Kyocera isn't suitable - I can see a schematics with a lens only :-) Is there a way to split the page (into 2 or 4 parts) and ptint parts separately (saving a scale, of course)?
Re: [gentoo-user] Time to move on?
On Sun, 22 Mar 2009 02:58:57 -0400, Philip Webb wrote: Now I'm doing an emerge -u world. I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I hope you use --oneshot every time or your world file will be a complete mess by now :( I also have a list of all the pkgs I have installed with dates + deps, which I keep upto-date by hand as I emerge items. I've never understood why 'emerge world' is considered standard: repeatedly, there are appeals for help here resulting from its shortcomings One or two problems a week against the thousands of people running it each day does not indicate a problem. I'd say that avoiding blockers etc by selectively skipping upgrades is more likely to lead to problems later. -- Neil Bothwick Windows 98, the most installed system in the world, I know, I've done it 5 or 6 times myself. signature.asc Description: PGP signature
Re: [gentoo-user] Time to move on?
On Sunday 22 March 2009 07:50:04 Mike Diehl wrote: So here is the question: Are these just growing pains, or is this the trend with Gentoo? If I resolve to update frequently, will these problems become more rare? I've been using Gentoo for 4 years now, my main desktop is still running code that I compiled on the first install in 2005. And I'm on my third Gentoo notebook in a row. Absurb update issues simply don't happen, as long as you follow the rules: Update weekly on ~arch Update monthly on arch Adjust to suit your needs. You ran into the mktemp issue, which feels about a year old from this corner, so Iguess you have not been updating regularly. I'm not sure where you got the advice to update only when you need a new feature or a fix, but it is not workable in practice. Gentoo does have issues, but the majority of them are with changes to packages, not changes to Gentoo. Remember that with Gentoo you are rebuilding a live system on the actual system itself. We don't have build farms that rebuild the entire distro and push out new rpms nightly - so the problems that can hit Gentoo don't happen to binary distro users. Take expat. It got an upgrade a long time ago which coudl break Gnome entirely if you didn't do it right. There was nothing the devs could do really, because that's how those packages were written. The normal case is to have a bare machine, build expat, then build Gnome. On Gentoo, you want to do all of this while using the Gnome that needs to be rebuilt. If you update regularly, you'll find lots of people around who know what the steps are and can help. Today, most of us have forgotten and need to turn to Google to find the howtos. If you find this happens more and more often with gentoo, it is probably a symptom of more and more useful packages out there, that are being developed faster with more features. See it as a sign of success on the part of FLOSS rather than a failing of Gentoo. And I would advise AGAINST gentoo on your out-of-state servers. They need too much pampering to keep them stable. I have about 100 machines at work, a mixture of 30% FreeBSD, SuSE ugh, some Centos and even a single lone Solaris machine. The worst of the lot has to be the SVN server, running Gentoo. No-one will touch it anymore, and the last time I did, I broke it horribly with a conflict between portage and cpan Perl modules. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Re: extending /usr partition...
On Sunday 22 March 2009 08:39:20 Francesco Talamona wrote: On Saturday 21 March 2009, Alan McKinnon wrote: It's correct, and it also highlights just what a PITA it is to manipulate traditional disk partitions. With lvm, this becomes a breeze. With ZFS (we might see it one day) this becomes invisible. I thought it was already there: sys-fs/zfs-fuse Yes, there is a fuse module. I meant in-kernel -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Time to move on?
On Sat, 21 Mar 2009 23:50:04 -0600, Mike Diehl wrote: However, lately, Gentoo seems to have been plagued with problems. Circular blockers. 32/64 bit libraries. Package re-organization. Others. That's inevitable with a versionless distro like Gentoo. With the other distros you have mentions, when the relationship between two packages changes, you don't notice because you only make the switch with what is effectively a re-install. With Gentoo,it is possible to try to upgrade one of the packages without the other, hence the need for blockers. So here is the question: Are these just growing pains, or is this the trend with Gentoo? If I resolve to update frequently, will these problems become more rare? It is a reducing trend, as portage gets better at resolving things automatically. The current method of resolving blockers has greatly reduced the time spent on them compared with a year ago. Frequent updates are a definite advantage, because when such issues do occur, they happen one at a time, making resolution much easier (like the mktemp which only needed a quick emerge -C then carry on). When you save up several months' worth of minor issues and let them all hit at once, it becomes more of a hassle to sort out. You should also run emerge --sync followed by glsa-check at least once a week to make sure you don't miss out on important security fixes. Another reason to keep up to date. -- Neil Bothwick Bother, said Pooh, as he drained the vodka bottle dry. signature.asc Description: PGP signature
Re: [gentoo-user] extending /usr partition...
On Saturday 21 March 2009 23:13:49 BRM wrote: So, unless you are looking to use LVM in a soft-RAID solution between multiple physical drives, not multiple partitions on the same drive, (e.g. partition A = sda1 + sda2, with mirror on sdb1+sdb2), then I would not suggest it as should anything happen, it'll make data recovery that much harder. LVM does not and should not provide data integrity features. You lost a drive. The data on it goes away. What did you expect would happen? That the data on it would magically reconstruct itself? In a situation like that, losing a drive with LVM is only slightly more inconvenient (one or two more steps) than losing the same drive without LVM (which is horribly inconvenient by itself). Please don't blame LVM for what is actually a user error. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Re: extending /usr partition...
On Sunday 22 March 2009 08:36:31 Francesco Talamona wrote: With or without LVM if you lose a drive then you've lost the data on it. LVM does have the capability of assembling a partially damaged volume group just not a partially damaged logical volume which, when you think about it, makes sense. And you can also throw in the standard warning about backing up your data. The point is that LVM adds an extra layer of complexity. Apparently you have not considered the enormous complexity inside the drive itself. The added complexity of LVM is tiny in comparison to what goes on there. LVM adds flexibility at the cost of one more thing to think about. You should already be performing backups and have redundancy plans (keeping in mind that it is a sheer miracle of modern science that the drive even works at all) For the occasional case where LVM does fail you. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Time to move on?
Hi, you also have the chance of running emerge -DuavN system. That way you can be sure that your system is stable without updating every program you might only need once in a blue moon or you are allready seticfied with. I would allways have an eye on the GLSA. You can do this in the forum, with rss or use something like glsa-check -t all kh
Re: [gentoo-user] extending /usr partition...
Am Samstag, 21. März 2009 19:39:08 schrieb Jarry: I remember having lvm2 a few years ago, and despite of that I could not extend any partition, which was being used. What is then lvm2 good for, if I can not extend partitions on-the-fly? I can not unmount /usr before extending... The filesystem has to support resizing. A few years ago ext3 only supported offline resize. That has changed meanwhile. Bye... Dirk signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: [OT] splitting and printing big image
On Sunday 22 March 2009, Andrew Gaydenko wrote: Hi! I have a PDF document. One of documet's page is a (scalable) schematics. Printing the page via my A4 Kyocera isn't suitable - I can see a schematics with a lens only :-) Is there a way to split the page (into 2 or 4 parts) and ptint parts separately (saving a scale, of course)? Maybe you can convert the page with pdftk to an imagem and then gimp it. Ciao Francesco -- Linux Version 2.6.28-gentoo-r3, Compiled #1 SMP PREEMPT Sun Mar 8 12:38:59 CET 2009 Two 1GHz AMD Athlon 64 Processors, 4GB RAM, 4018.04 Bogomips Total aemaeth
[gentoo-user] Seamonkey, saving files after updating gtk+
Hi, I'm not sure if the gtk+ update is relevant here but it was the only thing I could find that was recently upgraded that may fit. It used to be that if I was saving a file, picture or attachment in Seamonkey and created a new folder, it would enter the new folder when I created it without me having to double click it. Right now, I right click and select 'save file as' or 'save image as' and a pop up appears as usual. I then click 'create folder' and it makes a space in the directory for me to type in the name. I do that then hit return and it creates the new folder but leaves me in the old folder that I was just in. Example, I'm in /home/dale/Desktop and ask it to create /home/dale/Desktop/foo, the old way puts me in the foo directory. The new way leaves me in /home/dale/Desktop. I then have to double click the foo entry to enter it to save my file. I'm so used to the old way that I sometimes forget and end up with a lot of stuff on my desktop with no idea sometimes what folder it should be in. This is a list of the packages that were upgraded according to genlop: Fri Mar 20 04:58:16 2009 sys-apps/portage-2.2_rc26 Fri Mar 20 04:58:16 2009 sys-apps/portage-2.2_rc26 Fri Mar 20 04:58:51 2009 gnome-base/gnome-common-2.24.0 Fri Mar 20 05:03:58 2009 dev-libs/libxml2-2.7.3 Fri Mar 20 05:04:53 2009 dev-libs/libgamin-0.1.10-r2 Fri Mar 20 05:06:06 2009 media-libs/audiofile-0.2.6-r4 Fri Mar 20 05:07:19 2009 x11-libs/pixman-0.12.0 Fri Mar 20 05:08:56 2009 x11-libs/cairo-1.8.6-r1 Fri Mar 20 05:10:24 2009 dev-java/gjdoc-0.7.9-r1 Fri Mar 20 05:11:22 2009 app-text/iso-codes-3.6 Fri Mar 20 05:11:34 2009 x11-misc/icon-naming-utils-0.8.7 Fri Mar 20 05:13:18 2009 x11-themes/gnome-icon-theme-2.24.0 Fri Mar 20 05:18:05 2009 dev-libs/glib-2.18.4-r1 Fri Mar 20 05:20:26 2009 x11-libs/pango-1.22.4 Fri Mar 20 05:20:59 2009 dev-libs/atk-1.24.0 Fri Mar 20 05:22:17 2009 net-libs/libsoup-2.24.3 Fri Mar 20 05:23:10 2009 dev-libs/libcroco-0.6.2 Fri Mar 20 05:24:20 2009 dev-python/pygobject-2.16.1 Fri Mar 20 05:40:42 2009 x11-libs/gtk+-2.14.7-r2 Fri Mar 20 05:42:35 2009 gnome-base/gconf-2.24.0 Fri Mar 20 05:42:46 2009 gnome-base/gail-1000 Fri Mar 20 05:44:10 2009 x11-libs/libwnck-2.24.2 Fri Mar 20 05:45:49 2009 gnome-base/gnome-keyring-2.22.3-r1 Fri Mar 20 05:48:19 2009 dev-python/pygtk-2.14.0 Fri Mar 20 05:49:32 2009 gnome-extra/libgsf-1.14.11 Fri Mar 20 05:51:12 2009 gnome-base/librsvg-2.22.3 Fri Mar 20 05:56:03 2009 gnome-base/gnome-vfs-2.24.0 Fri Mar 20 05:57:01 2009 gnome-base/libgnome-2.24.1 Fri Mar 20 06:00:45 2009 gnome-base/libbonoboui-2.24.0 Fri Mar 20 06:03:40 2009 gnome-base/libgnomeui-2.24.0 Fri Mar 20 06:14:03 2009 gnome-extra/evolution-data-server-2.24.5-r2 Is this because of the gtk+ upgrade? If not, any idea what could cause this? If it is the upgrade, how do I go back to the old way when it goes into the folder automatically? Thanks for any ideas you may have. Dale :-) :-)
emerge-log (was: Re: [gentoo-user] Time to move on?)
Philip Webb schrieb am 22.03.2009 07:58: I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I also have a list of all the pkgs I have installed with dates + deps, which I keep upto-date by hand as I emerge items. I've never understood why 'emerge world' is considered standard: repeatedly, there are appeals for help here resulting from its shortcomings (was it copied from Free BSD when Gentoo was originally created ? ). But before I could do that, I had to upgrade portage, with made sense. When I go to emerge -u portage, I'm told: sys-apps/mktemp (is blocking sys-apps/coreutils-7.1) My hand-made list of pkgs tells me that I removed Mktemp 080419 that it had been required for Debianutils, of which my current version is 2.28.5 installed 090314 . Others had problems with this block, perhaps 1 year ago, so if you really are that far behind in updating, you should search the list archive to see what the advice was back then: IIRC a new version of Debianutils incorporated the Mktemp stuff, so they became incompatible. You know that such a list already exists. It is /var/log/emerge.log. To get useful information out of it app-portage/genlop comes handy. genlop -u mktemp * sys-apps/mktemp Sat Mar 3 18:14:30 2007 sys-apps/mktemp-1.5 Tue Dec 18 01:37:31 2007 sys-apps/mktemp-1.5 Sat Apr 12 23:15:42 2008 sys-apps/mktemp-1.5 Regards, Daniel signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] [OT] splitting and printing big image
On Sun, 22 Mar 2009 11:04:03 +0300, Andrew Gaydenko wrote: I have a PDF document. One of documet's page is a (scalable) schematics. Printing the page via my A4 Kyocera isn't suitable - I can see a schematics with a lens only :-) Is there a way to split the page (into 2 or 4 parts) and ptint parts separately (saving a scale, of course)? kdeprint has a poster print option. -- Neil Bothwick Suicide is the most sincere form of self-criticism. signature.asc Description: PGP signature
Re: [gentoo-user] extending /usr partition...
Am Samstag, 21. März 2009 19:39:08 schrieb Jarry: And one more counter-argument: with traditional partitions I can select where a certain partition is (physically). Those partitions accessed frequently I put to the beginning of the disk with higher transfer-rate. In my case, it makes quite difference: obelix ~ # hdparm -t /dev/md2 Timing buffered disk reads: 252 MB in 3.02 seconds = 83.23 MB/sec obelix ~ # hdparm -t /dev/md9 Timing buffered disk reads: 150 MB in 3.02 seconds = 49.72 MB/sec Who says you'd loose that ability with LVM? You could 1) Create two md devices and put one volume group on top of each calling them vg-fast and vg-slow, then put your logical volumes into either VG. 2) Keep your setup with 9 md devices and put one large VG on top, then for each LV tell LVM on which physical volume (PV) it should reside (don't know wether that really works, though. Just guessing that this is what the PhysicalVolumePath in lvcreate is meant for, the man page leaves it open). Bye... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] extending /usr partition...
Dirk Heinrichs wrote: I remember having lvm2 a few years ago, and despite of that I could not extend any partition, which was being used. What is then lvm2 good for, if I can not extend partitions on-the-fly? I can not unmount /usr before extending... The filesystem has to support resizing. A few years ago ext3 only supported offline resize. That has changed meanwhile. Are you sure? man resize2fs says: ...The resize2fs program will resize ext2 or ext3 file systems. It can be used to enlarge or shrink an unmounted file system located on device. If the filesystem is mounted, it can be used to expand the size of the mounted filesystem, assuming the kernel supports on-line resizing... If I understand correctly, you can only extend ext3-partition with it. But shrinking must be done off-line, on unmounted fs... Concerning resizing, I was wrong with blaming lvm2 for not being able to resize fs online. It is really problem of filesystem which might (or not) support it. I'm using ext3 because it is one of two linux-supported fs I have some experience with (the other is reiserfs, and I have some serious doubts about its future)... Jarry -- ___ This mailbox accepts e-mails only from selected mailing-lists! Everything else is considered to be spam and therefore deleted.
Re: [gentoo-user] Re: extending /usr partition...
[...] The point is that LVM adds an extra layer of complexity. I used LVM paired with soft RAID, and when I needed to boot from a liveCD I discovered that I had to rebuild the setup by hand. You mean the 1 extra command that's needed to assemble a VG?
[gentoo-user] Re: depclean wants to remove qt-4.4.2
Daniel Pielmeier schrieb: Marc Blumentritt schrieb am 21.03.2009 14:33: Hi, when I run emerge -p --depclen, I get these results: [...] These are the packages that would be unmerged: x11-libs/qt selected: 4.4.2 protected: none omitted: 3.3.8b-r1 x11-libs/qt-assistant selected: 4.4.2-r1 protected: none omitted: none x11-libs/qt-xmlpatterns selected: 4.4.2 protected: none omitted: none [...] Can someone explain me this? Yeah these packages are not needed by others anymore. If you really want/need them which i doubt you can put x11-libs/qt into the world file. Starting with qt-4.4 the ebuild has been split up into components. x11-libs/qt is just a meta ebuild nothing needs to depend on. Dependencies are set upon the components. So if you really want all qt stuff even if you do not need parts of it put x11-libs/qt in your world file. OK, thanks for making this clear for me. Marc
Re: [gentoo-user] Time to move on?
090322 Neil Bothwick wrote: On Sun, 22 Mar 2009 02:58:57 -0400, Philip Webb wrote: Now I'm doing an emerge -u world. I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I hope you use --oneshot every time or your world file will be a complete mess by now :( Yes, there's always someone who says that (grin). Of course, it's 2nd nature to check for 'W' or 'S' in my list if the pkg has neither, make sure to 'emerge -1 pkgname'. Anyway, isn't 'world' going to vanish with the new '@' sets ? I also have a list of all the pkgs I have installed with dates + deps, which I keep upto-date by hand as I emerge items. I've never understood why 'emerge world' is considered standard: repeatedly, there are appeals for help resulting from its shortcomings. One or two problems a week against the thousands of people running it each day does not indicate a problem. I'd say that avoiding blockers etc by selectively skipping upgrades is more likely to lead to problems later. No, people run 'emerge world' in the background, miss the messages then run into nasty trouble for omitting RR or 'etc-update'. That's the spirit of Ubuntu the rest, not the hands-on Gentoo approach. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Time to move on?
On Sat, 21 Mar 2009 19:17:53 -0600 Mike Diehl mdi...@diehlnet.com wrote: It seems that as long as I keep rebuilding machines from a current live CD, all is well. But if I try to upgrade anything else, I end up having to reformat. I've been using Gentoo long enough to have actually met Daniel Robbins in person, but I'm considering moving to a different distribution. I would say that if you do a complete world update at least every six months, followed by revdep-rebuild, keeping Gentoo up-to-date should be relatively painless, excluding all the blockers you have to resolve. ie.: emerge -uDNav world revdep-rebuild -i -- -a The libselinux problems you ran into are known, but that's also the reason why libselinux is masked on all recent profiles. /loki_val
Re: [gentoo-user] extending /usr partition...
On Sunday 22 March 2009 12:59:15 Jarry wrote: Are you sure? man resize2fs says: ...The resize2fs program will resize ext2 or ext3 file systems. It can be used to enlarge or shrink an unmounted file system located on device. If the filesystem is mounted, it can be used to expand the size of the mounted filesystem, assuming the kernel supports on-line resizing... If I understand correctly, you can only extend ext3-partition with it. But shrinking must be done off-line, on unmounted fs... Yes, and this is true for *any* filesystem. Some mature filesystems out there don't support resizing at all in any form. Enlarging a filesystems on line is easy. Make the device holding it bigger, add new inodes, add inodes to the inode tree and Hey Presto! filesystem is bigger. It's quick as you are *guaranteed* that the new inodes are not in use until the resize ends. Reducing is entirely different. You have to take inodes in use at the end of the filesystem, move them to somewhere else, fix the pointers in other inodes that point to them, repeat for all other inodes that will have to go away after the resize. Yuck. Tricky code :-) You are going to have this problem with any inode-based filesystem, not just ext3. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Time to move on?
On Sun, 22 Mar 2009 07:37:50 -0400, Philip Webb wrote: 090322 Neil Bothwick wrote: I hope you use --oneshot every time or your world file will be a complete mess by now :( Yes, there's always someone who says that (grin). I wouldn't want to disappoint you :) Of course, it's 2nd nature to check for 'W' or 'S' in my list if the pkg has neither, make sure to 'emerge -1 pkgname'. I just use -1 whenever I re-emerge anything, whether it's in world or not. Anyway, isn't 'world' going to vanish with the new '@' sets ? No. One or two problems a week against the thousands of people running it each day does not indicate a problem. I'd say that avoiding blockers etc by selectively skipping upgrades is more likely to lead to problems later. No, people run 'emerge world' in the background, miss the messages then run into nasty trouble for omitting RR or 'etc-update'. Don't blame hammers because people try to use them to drive screws. Letting idiots break their systems is a refreshing sign of Gentoo's lack of idiot-proofing :) -- Neil Bothwick All mail what i send is thoughly proof-red, definately! signature.asc Description: PGP signature
Re: [gentoo-user] [OT] splitting and printing big image
On Sunday 22 March 2009 12:57:15 Neil Bothwick wrote: On Sun, 22 Mar 2009 11:04:03 +0300, Andrew Gaydenko wrote: I have a PDF document. One of documet's page is a (scalable) schematics. Printing the page via my A4 Kyocera isn't suitable - I can see a schematics with a lens only :-) Is there a way to split the page (into 2 or 4 parts) and ptint parts separately (saving a scale, of course)? kdeprint has a poster print option. Aha, 'poster' CLI has helped me. Thanks!
[gentoo-user] Re: extending /usr partition...
On Sunday 22 March 2009, Albert Hopkins wrote: [...] The point is that LVM adds an extra layer of complexity. I used LVM paired with soft RAID, and when I needed to boot from a liveCD I discovered that I had to rebuild the setup by hand. You mean the 1 extra command that's needed to assemble a VG? It wasn't that easy, that's what I did in the end: 1) vgchange -a n 2) vgexport -a 3) vgimport -a 4) vgscan --mknodes 5) vgchange -a y Maybe 4) and 5) alone would do the trick... I don't remember which livecd I used then, except for the fact that I had to reboot with dolvm2 option; now I have a tested first aid kit with notes on paper (all my notes were on those discs at the time) with well known and proven liveCDs. When I have to resize/redesign my partitions I simply find easier rsync plus a reboot. I rsync the live system while I use it, than reboot to a liveCD to rsync the file changed meanwhile (to minimize downtime). So I shortly tossed LVM and since I live happily without. Ciao Francesco -- Linux Version 2.6.28-gentoo-r3, Compiled #1 SMP PREEMPT Sun Mar 8 12:38:59 CET 2009 Two 1GHz AMD Athlon 64 Processors, 4GB RAM, 4018.04 Bogomips Total aemaeth
Re: [gentoo-user] start X at startup without a login manager
I was thinking about the proper way to set-up the autologin some time
ago.
Some hints which came to my mind :
1) should be started as a service so restart/start/stop may be used and
doesn't leave an open root shell
2) should be restarted with ctrl+alt+backspace without losing the
keyboard focus (may be a problem when not using telinit)
3) try to avoid the numerous wrappers of login managers in /etc/X11
4) should be able to come back to login manager by modifying only one
conf.d file (or env.d ?)
--- so below is the way I have set it up for the moment :
I did a kind of merge of the xdm and startUS.sh (script found on the
gentoo forums) so I don't have x11-apps/xdm but I have a xdm init script
(which should have another name but it's just a modified version of the
original xdm script)
[xdm patch attached]
To avoid the unresponsive keyboard : I always use telinit so :
x:a:once:/etc/X11/startDM.sh
is appended in my /etc/inittab.
In startDM.sh a modification is done to make start-stop-daemon drop its
privs and set the minimum env needed by startx then xinit.
[startDM.sh patch attached]
Notice two facts : env X=y start-stop-daemon is used because I was not
able to use several --env options.
My user's .xinitrc sources its .bash_profile (which source /etc/profile)
to initialize the other variables.
(I would like to export the bash completion to my whole X session, but it's
another problem...)
An alternative is the make start-stop-daemon launch 'su -- -l'
but it's dirty because of the need to store the pid.
So the first problem is that the xdm script doesn't know the pid of xinit
because even without 'su', start-stop-daemon knows about startx, not
xinit.
The second one is that ctrl+alt+backspace isn't trapped correctly.
Should 'xinit restart' be the direct work of the daemon in the autologin
case ?
In the autologin case which imho implies the user has a .xinitrc, startx
is only useful for the 2 or 3 lines around mcookie.
Should startx be directly in /etc/init.d .
(as said in the header it's a old sample of this script) and the deep
meaning of runlevel (multiuser / graphic) should be think from the
beginning to understand the right way to organise the X11 launch stuff.
What about putting startx's $defaultserverarg and $enable_xauth in a
/etc/conf.d/xdm (or better : /etc/conf.d/xinit) ?
(the local.start is a hacky but short and understandable way to do
though :), a quick heavier case is there :
http://www.gentoo-wiki.info/TIP_Passwordless_Login)
Anyway, I'm still a bit lost in the quest of the cleanest way
from system init to ~/xinitrc.
So any comment, advice, whatever ... would be greatly appreciated.
Raph
On Fri, Mar 20, 2009 at 11:09:56PM +0800, fei huang wrote:
I don't have any xdm, gdm stuff but would like to start my windows
manager directly at startup, cause I'm the only one that use it.
here is my solution:
I use runlevel 3 as default, and add a line of code in
/etc/conf.d/local.start:
su - myname -c startx
this works just fine except my scim panel would not shown as before, but if
I login in normally with my user name, and type startx manually,
everything works perfect. I'm wondering what is the difference with those
two steps that cause the problem,
ps shows the scim processes are just running normally, for reference, I
pasted my xinitrc here:
export XMODIFIERS='@im=SCIM'
export GTK_IM_MODULE=scim
scim -d
xrdb -merge ~/.Xresources
urxvtd -q -f -o
conky -q
exec awesome
any ideas?
thanks
fei
--- xdm.old 2009-02-06 12:11:29.0 +0100
+++ xdm 2009-02-06 12:15:53.0 +0100
@@ -85,6 +85,12 @@
EXE=/usr/bin/wdm
PIDFILE=
;;
+ none)
+ test -n $(id -u ${XUSER} 2/dev/null) \
+ EXE=/usr/bin/startx \
+ PIDFILE=/var/run/x.pid \
+ NAME=startx
+ ;;
*)
EXE=
# Fix #65586, where MY_XDM is empty so EXE=somedir
@@ -140,6 +146,7 @@
save_options service ${EXE}
save_options name${NAME}
save_options pidfile ${PIDFILE}
+ save_options xuser${XUSER}
if [ -n ${CHECKVT-y} ] ; then
if vtstatic ${CHECKVT:-7} ; then
@@ -154,7 +161,11 @@
fi
fi
- /etc/X11/startDM.sh
+ if [ -n ${XUSER} ] [ -x /sbin/telinit ]; then
+ telinit a /dev/null 21
+ else
+ /etc/X11/startDM.sh
+ fi
eend 0
}
--- startDM.sh.old 2009-03-22 00:33:04.0 +0100
+++ startDM.sh 2009-03-22 00:34:32.0 +0100
@@ -14,17 +14,27 @@
[ -r ${svclib}/sh/rc-services.sh ] . ${svclib}/sh/rc-services.sh
fi
-# Great new Gnome2 feature, AA
-# We enable this by default
-export GDK_USE_XFT=1
export SVCNAME=xdm
EXEC=$(get_options service)
NAME=$(get_options name)
Re: [gentoo-user] Time to move on?
Am Sonntag, 22. März 2009 02:17:53 schrieb Mike Diehl: Ok now, I'm getting fed up with all of the breakage that I've seen in Gentoo in the last few months. I'm trying to upgrade MythTV. Emerge told me to upgrade my profile, which I did. Now I'm doing an emerge -u world. But before I could do that, I had to upgrade portage, with made sense. When I go to emerge -u portage, I'm told: sys-apps/mktemp (is blocking sys-apps/coreutils-7.1) So I do: emerge -C mktemp Depending on which package you unmerge before the blocker is resolved, this can break your system. This is while paludis and newer versions of portage can be told to resolve blockers automatically. Has Gentoo become such a moving target that it's no longer suitable for normal, every day, usage? Not if you keep it up to date. Then, even a ~arch system is relatively painless to use. I've got another machine that needs to be upgraded, but hasn't been upgraded in some time. So it's profile is obsolete and many of the core packages have been moved around so much that there is no upgrade path from where it is now, to where Gentoo is. There always is. However, the longer the path, the more pitfalls may be on it. Is it time to start looking for a new distribution? It seems that as long as I keep rebuilding machines from a current live CD, all is well. ??? I don't understand. But if I try to upgrade anything else, I end up having to reformat. That also doesn't make sense. I've been using Gentoo long enough to have actually met Daniel Robbins in person, but I'm considering moving to a different distribution. dito. Remember, all I want to do is upgrade MythTV. So, then just do it. People here will help you to resolve the issues. Bye... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Time to move on?
On Sun, 2009-03-22 at 12:44 +0100, Peter Alfredsen wrote: I would say that if you do a complete world update at least every six months, followed by revdep-rebuild, keeping Gentoo up-to-date should be relatively painless, excluding all the blockers you have to resolve. ie.: emerge -uDNav world revdep-rebuild -i -- -a I've done this on a machine I hadn't touched in over 6 months. And, surprisingly, I was relieved that it came out fine. Though I did have the advantage of: * Having another machine that I upgrade regularly and so know what to look out for * I read and react to, if necessary, the elog messages from the ebuild chatter (I have them sent to my mailbox) * Checking the bug database if I run into a snag * General experience on how to maintain a (Gentoo) system That and, if a particular version does not work out for you, you often can downgrade to an older version. Or if you don't like the way something is built, even with the available USE flags, you can usually keep a simple patch and keep your own version in a private overlay. I love this stuff. This is why I use Gentoo. Contrast with another distro I use. I recently upgraded to version n+1 and am encountering all kinds of problems. I have versions of software installed that don't work or don't work the way they used to, but I can't go back. I can't install the older versions of packages because they depend on older versions of libs that no longer exist on version n +1 (and there are no such thing as SLOTs and revdep-rebuild). And even if I thought about downgrading the entire distro to version n that pretty much means a re-install of the entire OS (and then a re-update of the downgraded OS). I've submitted two bugs for version n+1 but one that I submitted in January hasn't even been responded to and the other was quickly closed as a WONTFIX. Not to criticize other distros (which is one reason why I didn't even name it), but my point is that they all have their pluses and minuses. For me at least, Gentoo comes with fewer minuses and when they do come they are usually easier to fix/get fixed. The caveat is that you actually have to know/care what you're doing.
Re: [gentoo-user] Re: extending /usr partition...
On Sun, 2009-03-22 at 13:53 +0100, Francesco Talamona wrote: You mean the 1 extra command that's needed to assemble a VG? It wasn't that easy, that's what I did in the end: 1) vgchange -a n 2) vgexport -a 3) vgimport -a 4) vgscan --mknodes 5) vgchange -a y #5 is all I've ever had to do. The first 3 look pointless to me. #5 takes care of #4 for you. Maybe 4) and 5) alone would do the trick... I don't remember which livecd I used then, except for the fact that I had to reboot with dolvm2 option; now I have a tested first aid kit with notes on paper (all my notes were on those discs at the time) with well known and proven liveCDs. I've never had to dolvm2 either. I'm guessing that's a Gentoo live cd thing. I rarely use the Gentoo live cds because they always seem out of date (although I understand they build daily snapshots now).
Re: [gentoo-user] resolving this block
On 21/03/09 Nick Fortino said: Hmm, that's odd. gtk+ does flag the fact that it includes gail with the block you are running into, and gail-1000 is used to make programs which depend on gail happy. emerge -uDN world should really just take care of this, as gail should be upgraded to gail-1000. At the end of the day, Daniel is right, unmerging gail should resolve the block, and everything should work when you are done upgrading. Portage should (and did for me) take care of this on it's own though. Hmm. Is it possible that I don't have the latest portage? I have sys-apps/portage-2.1.6.7. emerge --pretend --update sys-apps/portage shows nothing so I guess I have the latest one... Mike -- Michael P. Soulier msoul...@digitaltorque.ca Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction. --Albert Einstein pgpkTZVqAvC6X.pgp Description: PGP signature
Re: [gentoo-user] resolving this block
On Sun, 22 Mar 2009 11:35:56 -0400 Michael P. Soulier msoul...@digitaltorque.ca wrote: On 21/03/09 Nick Fortino said: Hmm. Is it possible that I don't have the latest portage? I have sys-apps/portage-2.1.6.7. emerge --pretend --update sys-apps/portage shows nothing so I guess I have the latest one... echo sys-apps/portage /etc/portage/package.unmask echo sys-apps/portage /etc/portage/package.keywords I have 2.2_rc26, umasked 2.2 from the beginning and have yet to face any problem with it (unlike with 2.1, which seem to figure in quite a few threads here) :) -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
[gentoo-user] Gentoo as OpenVZ/Virtuozzo guest
Hi list! I recently acquired a virtual server at hosteurope.de However, I slowly loose my patience with the OpenSuse installation and would like to have my fellow Gentoo on it. I already found this guide: http://www.gentoo.org/proj/en/vps/openvz-howto.xml But it doesn't say much about how to run the guest. How do I have to configure the kernel? Which sources can I use? What about a hardened system? What is this vzfs file system? Can I use partitions? When anything goes wrong, can I still use the repair mode? How does it shut down my server before giving me a fresh installation from which I can chroot into it? Thanks in advance, Florian Philipp signature.asc Description: OpenPGP digital signature
[gentoo-user] setting a USE flags for many packages
Hi, I normally like to have the documentation built for all packages. Therefore I have added 'doc' to /etc/make.conf USE But I don't need documentation for the packages in kde-base/* and kde-misc/*. Since some of these take over an hour just for doxygen (on a Phenom II 3GHz) I'd like to disable the generation of documentation for all these packages. How can I set the '-doc' use flags for all packages matching kde-base/* and kde-misc/* (There are hundreds of them!) Note, I cannot use 'USE=-doc' on the command line since emerge --update --complete-graph --reinstall changed-use world would then reinstall many other packages to remove the documentation which I doesn't want. Many thanks for a hint, Helmut Jarausch Lehrstuhl fuer Numerische Mathematik RWTH - Aachen University D 52056 Aachen, Germany
Re: [gentoo-user] can't make rt2570 module FIXED
--- On Fri, 3/13/09, maxim wexler bliss...@yahoo.com wrote: From: maxim wexler bliss...@yahoo.com Subject: [gentoo-user] can't make rt2570 module To: gentoo-user@lists.gentoo.org Received: Friday, March 13, 2009, 2:32 PM Hi group, I found a bug report for this but that involved a problem with the kernel config. In this case, portage doesn't seem to have a problem with that. This is for a NovaTech usb-wifi gizmo. ID: 0eb0:9020 If you're having trouble with your Ralink product go here: rt2x00.serialmonkey.com __ Ask a question on any topic and get answers from real people. Go to Yahoo! Answers and share what you know at http://ca.answers.yahoo.com
Re: [gentoo-user] Gentoo as OpenVZ/Virtuozzo guest
Florian Philipp wrote: I recently acquired a virtual server at hosteurope.de ... How do I have to configure the kernel? Which sources can I use? I'd say you do not need any sources, because as you wrote, you have guest. openvz/virtuozzo uses one single kernel running on host (similar as vserver-linux). As it is common, company offering virtual servers has a pool of guest-systems you can choose from (iirc, hosteurope offers debian, ubuntu, suse, and some kinds of windoze). You can not install a guest system of your choice, but you can ask them and maybe they will prepare it for you. Then you get root access to your guest system, and you can do with it whatever you want (except of building new kernel)... Jarry -- ___ This mailbox accepts e-mails only from selected mailing-lists! Everything else is considered to be spam and therefore deleted.
Re: [gentoo-user] Gentoo as OpenVZ/Virtuozzo guest
Jarry schrieb: Florian Philipp wrote: I recently acquired a virtual server at hosteurope.de ... How do I have to configure the kernel? Which sources can I use? I'd say you do not need any sources, because as you wrote, you have guest. openvz/virtuozzo uses one single kernel running on host (similar as vserver-linux). As it is common, company offering virtual servers has a pool of guest-systems you can choose from (iirc, hosteurope offers debian, ubuntu, suse, and some kinds of windoze). You can not install a guest system of your choice, but you can ask them and maybe they will prepare it for you. Then you get root access to your guest system, and you can do with it whatever you want (except of building new kernel)... Jarry Never mind, I've just found this howto: http://log.onthebrink.de/2008/04/gentoo-on-1-vserver.html Can't promise that I won't come back with more questions, though ;) signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] setting a USE flags for many packages
On Sunday 22 March 2009 18:57:16 Valmor de Almeida wrote: Helmut Jarausch wrote: Hi, ... How can I set the '-doc' use flags for all packages matching kde-base/* and kde-misc/* (There are hundreds of them!) I don't know whether the /etc/portage/package.use file will accept this syntax but maybe you can do something like kde-base/* -doc It's not supported: --- Invalid atom in /etc/portage/package.use: kde-base/* The OP has two choices really: remove the flags from USE and explicitly list the packages for which he does want documentation, or insert the flag into USE and explictly list the packages for which he wants the documentation removed. So you are reduced to grep, sed, awk and friends. Here's one (grossly inefficient) method: equery -q hasuse doc | xargs qatom | cut -f1-2 -d ' ' | sed -e 's/ /\//' -e 's/$/ -doc/' /etc/portage/package.use/package.use -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] resolving this block
On Sunday 22 March 2009 17:35:56 Michael P. Soulier wrote: On 21/03/09 Nick Fortino said: Hmm, that's odd. gtk+ does flag the fact that it includes gail with the block you are running into, and gail-1000 is used to make programs which depend on gail happy. emerge -uDN world should really just take care of this, as gail should be upgraded to gail-1000. At the end of the day, Daniel is right, unmerging gail should resolve the block, and everything should work when you are done upgrading. Portage should (and did for me) take care of this on it's own though. Hmm. Is it possible that I don't have the latest portage? I have sys-apps/portage-2.1.6.7. emerge --pretend --update sys-apps/portage shows nothing so I guess I have the latest one... You have the latest *stable* portage. There are others in the unstable branch with the latest cutting-edge features. Lucky for us, Zac has done an excellent job in the last year of stopping just short of the bleeding edge -- alan dot mckinnon at gmail dot com
[gentoo-user] kismet fatal error
Hi group, Now that rt2570 module is installed I'd like to find out if there's a signal or if the Ralink device even works. iwconfig finds: Link Quality=0/100 Signal level:-120 dBm Noise level:-93 dBm This seems to say there is no signal. Right? Here's kismet heat...@kyzyl ~ $ kismet_client FATAL: Could not connect to localhost:2501. Google doesn't help. Changing /etc/hosts line 127.0.0.1 heathen localhost to the more generic 127.0.0.1 localhost.localdomain localhost doesn't work. Maxim __ Get the name you've always wanted @ymail.com or @rocketmail.com! Go to http://ca.promos.yahoo.com/jacko/
RE: [gentoo-user] Time to move on?
-Original Message- From: Philip Webb [mailto:purs...@ca.inter.net] Sent: March 22, 2009 7:38 AM To: gentoo-user@lists.gentoo.org Subject: Re: [gentoo-user] Time to move on? 090322 Neil Bothwick wrote: On Sun, 22 Mar 2009 02:58:57 -0400, Philip Webb wrote: Now I'm doing an emerge -u world. I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I hope you use --oneshot every time or your world file will be a complete mess by now :( Yes, there's always someone who says that (grin). Of course, it's 2nd nature to check for 'W' or 'S' in my list if the pkg has neither, make sure to 'emerge -1 pkgname'. Anyway, isn't 'world' going to vanish with the new '@' sets ? I also have a list of all the pkgs I have installed with dates + deps, which I keep upto-date by hand as I emerge items. I've never understood why 'emerge world' is considered standard: repeatedly, there are appeals for help resulting from its shortcomings. One or two problems a week against the thousands of people running it each day does not indicate a problem. I'd say that avoiding blockers etc by selectively skipping upgrades is more likely to lead to problems later. No, people run 'emerge world' in the background, miss the messages then run into nasty trouble for omitting RR or 'etc-update'. That's the spirit of Ubuntu the rest, not the hands-on Gentoo approach. I run emerge world in the background and still run etc-update when I'm told to. Crontab emails plus portage's --quiet flag are awesome. Speaking of that, going to run that now.
Re: [gentoo-user] setting a USE flags for many packages
On Sun, 22 Mar 2009 17:43:29 +0100 (CET), Helmut Jarausch wrote: I normally like to have the documentation built for all packages. Therefore I have added 'doc' to /etc/make.conf USE Are you aware that the doc USE flag normally controls extra documentation, such as API docs for devs? Normal documentation, like man and info pages, is generally built by default. You're probably better off with -doc in make.conf and only enable it for the packages you need. -- Neil Bothwick WinErr 009: Horrible bug encountered - God knows what has happened signature.asc Description: PGP signature
Re: [gentoo-user] Time to move on?
Philip Webb wrote: 090322 Neil Bothwick wrote: On Sun, 22 Mar 2009 02:58:57 -0400, Philip Webb wrote: Now I'm doing an emerge -u world. I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I hope you use --oneshot every time or your world file will be a complete mess by now :( Yes, there's always someone who says that (grin). Of course, it's 2nd nature to check for 'W' or 'S' in my list if the pkg has neither, make sure to 'emerge -1 pkgname'. Anyway, isn't 'world' going to vanish with the new '@' sets ? Nope, they are still there and the sets are working too. Dale :-) :-)
Re: [gentoo-user] Time to move on?
Philip Webb wrote: 090322 Neil Bothwick wrote: On Sun, 22 Mar 2009 02:58:57 -0400, Philip Webb wrote: Now I'm doing an emerge -u world. I never do that: I always do 'emerge -Dup world', then decide which packages to update emerge them individually. I hope you use --oneshot every time or your world file will be a complete mess by now :( Yes, there's always someone who says that (grin). Of course, it's 2nd nature to check for 'W' or 'S' in my list if the pkg has neither, make sure to 'emerge -1 pkgname'. Anyway, isn't 'world' going to vanish with the new '@' sets ? Nope, they are still there and the sets are working too. Dale :-) :-)
Re: [gentoo-user] setting a USE flags for many packages
Thanks for the albeit sad news! Helmut. -- Helmut Jarausch Lehrstuhl fuer Numerische Mathematik RWTH - Aachen University D 52056 Aachen, Germany
Re: [gentoo-user] setting a USE flags for many packages
Neil Bothwick wrote: On Sun, 22 Mar 2009 17:43:29 +0100 (CET), Helmut Jarausch wrote: I normally like to have the documentation built for all packages. Therefore I have added 'doc' to /etc/make.conf USE Are you aware that the doc USE flag normally controls extra documentation, such as API docs for devs? Normal documentation, like man and info pages, is generally built by default. You're probably better off with -doc in make.conf and only enable it for the packages you need. This is true. I have had -doc in make.conf for a long time and I still have all the man and info pages. I can't say that I have even missed anything really. Dale :-) :-)
[gentoo-user] Install on a Dell E4300
Hi all, I was trying to install Gentoo 2008.0 in a Dell E4300 but the kernel does not have support for the Ethernet controller: Intel Corporation 82567LM Gigabit Network Connection. Does anybody try to install Gentoo in that notebook? or does anybody knows a better way to get gentoo working on this dell notebook? I did the following : Get the stages 2 files, download the kernel 2.6.28 and compile it and after that I got support for my ethernet card and wireless card but the resto of the installation was really problematic, so I decided to make that question in these lists. thanks in advance and best regards. -- Francisco Rivas http://www.vaslibre.org.ve - And on the seventh day God said :wq and then make http://beck3r.wordpress.com/ Linux User (New) : #448324 Linux Machine (New) : 355187
Re: [gentoo-user] setting a USE flags for many packages
On Sunday 22 March 2009 21:53:25 Dale wrote: Neil Bothwick wrote: On Sun, 22 Mar 2009 17:43:29 +0100 (CET), Helmut Jarausch wrote: I normally like to have the documentation built for all packages. Therefore I have added 'doc' to /etc/make.conf USE Are you aware that the doc USE flag normally controls extra documentation, such as API docs for devs? Normal documentation, like man and info pages, is generally built by default. You're probably better off with -doc in make.conf and only enable it for the packages you need. This is true. I have had -doc in make.conf for a long time and I still have all the man and info pages. I can't say that I have even missed anything really. Same here. Well, almost. I have USE=doc set for 13 packages thinking at the time they were extremely useful. In practise the only one I have ever actually referred to is the html version of the mplayer man page. Go figure :-) -- alan dot mckinnon at gmail dot com
[gentoo-user] Resizing physical volume for lvm.
My current setup is: Device Boot Start End Blocks Id System /dev/sda1 * 1289423246023+ 83 Linux /dev/sda228953381 3911827+ 82 Linux swap /Solaris /dev/sda33382 24804 172080247+ 83 Linux /dev/sda4 24805 3040144957902+ 83 Linux where sda3 is an lvm volume and sda4 is free space. I'd like to to merge sda3 and sda4 into a single partition without losing the data on it, but I'm not sure if it is possible. My guess is that I can use fdisk to delete sda4 and sda3, create a sda3 partition starting at 3382 and ending at 30401, then use pvresize to enlarge it. This is from man pvresize: Expand the PV on /dev/sda1 after enlarging the partition with fdisk: pvresize /dev/sda1 Is that going to work or I'm going to lose all the data? P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4, and to do so I will need to split it in two separate partitions (/boot using ext3 and / using ext4). This way I'm not going to need extended partitions. --- TopperH http://topperh.blogspot.com signature.asc Description: PGP signature
Re: [gentoo-user] Time to move on?
On Sat, Mar 21, 2009 at 6:17 PM, Mike Diehl mdi...@diehlnet.com wrote: Ok now, I'm getting fed up with all of the breakage that I've seen in Gentoo in the last few months. Understood and personally felt. SNIP emerge -C mktemp Generally a *very* bad move unless you are *absolutely* sure that what you are removing is not needed to keep the system working. SNIP Has Gentoo become such a moving target that it's no longer suitable for normal, every day, usage? OK, I'm a putz who has used Gentoo now for (I think) 8-9 years. I'm not a developer, a programmer or a sys admin. Keep that in mind. Gentoo goes through phases of relative stability interrupted by periods of time where major problems dominate. (As seen by users, not the Lords of Gentoo (LoG)) Personally I think we're in one of those unfortunate periods of time where there is a relatively high number of issues. I'm seeing it on all my machines. It's taking far more of my time to deal with this than I wish it would. 1) ntp-update problems at boot time. 2) emerge -DuN world building lots of packages that are already on the system but the LoG has apparently changed flags so emerge wants to rebuild them. 3) New and unclear (to me) messages about portage flag overrides caused by overlays I've been using for a while. Am I frustrated like you? Yep. Very much so. Am I considering using something else? Quite a few thoughts. Do I think there's a better distro? Not that I know of. My workload: 1) Gentoo 64-bit desktop 2) Gentoo 32-bit desktop 3) Gentoo 32-bit mythbackend 4) Gentoo 32-bit mythfrontend 5) Gentoo 32-bit mythfrontend Cheers, Mark
Re: [gentoo-user] Resizing physical volume for lvm.
Hi, Momesso Andrea a écrit : My current setup is: Device Boot Start End Blocks Id System /dev/sda1 * 1289423246023+ 83 Linux /dev/sda228953381 3911827+ 82 Linux swap /Solaris /dev/sda33382 24804 172080247+ 83 Linux /dev/sda4 24805 3040144957902+ 83 Linux where sda3 is an lvm volume and sda4 is free space. You may consider setting your partition type to 8E (Linux LVM), which would give a better labeling of your table. I understand what you want to do, and I would process the same way but I'd rather make a backup before (of your partition table and data). As I never did it, I prefer not telling you to go ahead ;) Sincerely, Jil
Re: [gentoo-user] Resizing physical volume for lvm.
On Sunday 22 March 2009 22:15:14 Momesso Andrea wrote: My current setup is: Device Boot Start End Blocks Id System /dev/sda1 * 1289423246023+ 83 Linux /dev/sda228953381 3911827+ 82 Linux swap /Solaris /dev/sda33382 24804 172080247+ 83 Linux /dev/sda4 24805 3040144957902+ 83 Linux where sda3 is an lvm volume and sda4 is free space. I'd like to to merge sda3 and sda4 into a single partition without losing the data on it, but I'm not sure if it is possible. My guess is that I can use fdisk to delete sda4 and sda3, create a sda3 partition starting at 3382 and ending at 30401, then use pvresize to enlarge it. Correct. That's all there is to it. This is from man pvresize: Expand the PV on /dev/sda1 after enlarging the partition with fdisk: pvresize /dev/sda1 Is that going to work or I'm going to lose all the data? Your data is safe if you do exactly the steps you said above. Caveat: I have no idea why this doesn't work, but if you make sda4 an extended partition and create sda5 as a logical with exactly the same start and end as you describe above, you do in fact lose all data. Obviously there is a difference between a physical and a logical partition with the same location, but I don't know why this is. Which is a pity, as 4 logical partitions is a little too constrictive, I prefer the extra freedom to move things around with extended partitions. P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4, and to do so I will need to split it in two separate partitions (/boot using ext3 and / using ext4). This way I'm not going to need extended partitions. ext3 on /boot is pointless. The ext3 metadata takes up a considerable chunk of the space on a typical /boot, for no good reason at all - writes to it are exceptionally rare so there's no real-worlld benefit to the journal. Ext2 is ideal for /boot. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Resizing physical volume for lvm.
Am Sonntag, 22. März 2009 21:15:14 schrieb Momesso Andrea: P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4 Why do want to do that? ext4 is just a couple of months old and there's no proof of stability whatsoever for it. Better try it with uncritical data partitions (portage tree, distfiles) first. Bye... Dirk signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Time to move on?
On Sun, 22 Mar 2009 13:23:38 -0700, Mark Knecht wrote: 2) emerge -DuN world building lots of packages that are already on the system but the LoG has apparently changed flags so emerge wants to rebuild them. Don't use --newuse, use --reinstall changed-use. -- Neil Bothwick Most software is about as user-friendly as a cornered rat! signature.asc Description: PGP signature
Re: [gentoo-user] Resizing physical volume for lvm.
On Sun, 22 Mar 2009 22:35:35 +0200 Alan McKinnon alan.mckin...@gmail.com wrote: On Sunday 22 March 2009 22:15:14 Momesso Andrea wrote: My current setup is: Device Boot Start End Blocks Id System /dev/sda1 * 1289423246023+ 83 Linux /dev/sda228953381 3911827+ 82 Linux swap /Solaris /dev/sda33382 24804 172080247+ 83 Linux /dev/sda4 24805 3040144957902+ 83 Linux where sda3 is an lvm volume and sda4 is free space. I'd like to to merge sda3 and sda4 into a single partition without losing the data on it, but I'm not sure if it is possible. My guess is that I can use fdisk to delete sda4 and sda3, create a sda3 partition starting at 3382 and ending at 30401, then use pvresize to enlarge it. Correct. That's all there is to it. This is from man pvresize: Expand the PV on /dev/sda1 after enlarging the partition with fdisk: pvresize /dev/sda1 Is that going to work or I'm going to lose all the data? Your data is safe if you do exactly the steps you said above. Good to know! In any case backups are available, but I prefer not to use them if not necessary. Caveat: I have no idea why this doesn't work, but if you make sda4 an extended partition and create sda5 as a logical with exactly the same start and end as you describe above, you do in fact lose all data. Obviously there is a difference between a physical and a logical partition with the same location, but I don't know why this is. Which is a pity, as 4 logical partitions is a little too constrictive, I prefer the extra freedom to move things around with extended partitions. P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4, and to do so I will need to split it in two separate partitions (/boot using ext3 and / using ext4). This way I'm not going to need extended partitions. ext3 on /boot is pointless. The ext3 metadata takes up a considerable chunk of the space on a typical /boot, for no good reason at all - writes to it are exceptionally rare so there's no real-worlld benefit to the journal. Ext2 is ideal for /boot. Thanks for the advice. Will be a problem for lvm if I add a partition before it? I mean, will I need to change any config files while lvm is gonna reside on sda4 instead of sda3? --- TopperH http://topperh.blogspot.com signature.asc Description: PGP signature
Re: [gentoo-user] Resizing physical volume for lvm.
On Sun, 22 Mar 2009 21:39:29 +0100 Dirk Heinrichs dirk.heinri...@online.de wrote: Am Sonntag, 22. März 2009 21:15:14 schrieb Momesso Andrea: P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4 Why do want to do that? ext4 is just a couple of months old and there's no proof of stability whatsoever for it. Better try it with uncritical data partitions (portage tree, distfiles) first. Bye... Dirk Don't want to do that right now, but I like to have things set up to create me less problems as possible when I will decide to do the migration. In any case I don't consider the root partition of my laptop to be really critical. I'm not planning any migration in /home, that will remain ext3. --- TopperH http://topperh.blogspot.com signature.asc Description: PGP signature
Re: [gentoo-user] Time to move on?
On Sun, 2009-03-22 at 13:23 -0700, Mark Knecht wrote: Personally I think we're in one of those unfortunate periods of time where there is a relatively high number of issues. I'm seeing it on all my machines. It's taking far more of my time to deal with this than I wish it would. 1) ntp-update problems at boot time. Hmm, I'm not having any ntp-update problems on my machines. Have you submitted a bug report or searched the bug database? Obviously this isn't happening for everyone so if the right people don't know about it then you can't expect it to get fixed. 2) emerge -DuN world building lots of packages that are already on the system but the LoG has apparently changed flags so emerge wants to rebuild them. But that's what -N does! That's what it's documented to do. RTFM. If you don't want that behavior then don't use -N. Either use --reinstall changed-use or don't use any USE-specific flags. Personally I just let portage re-install as it doesn't really change anything if you haven't changed your use flags. 3) New and unclear (to me) messages about portage flag overrides caused by overlays I've been using for a while. These are probably warnings about overlays overriding settings in the regular portage profile. Some overlays do this. It's just a fact of life. Again, if you don't want to deal with it then don't use overlays or at least choose overlays that don't do potentially bad things. It's not the Gentoo devs responsibility if you bring in 3rd-party overlays that change stuff. So with the possible exception of #1, these appear to be but it hurts when I do that problems, not issues with Gentoo stability.
Re: [gentoo-user] Resizing physical volume for lvm.
On Sun, 22 Mar 2009 20:49:36 + Neil Bothwick n...@digimed.co.uk wrote: On Sun, 22 Mar 2009 21:39:29 +0100, Dirk Heinrichs wrote: P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4 Why do want to do that? ext4 is just a couple of months old and there's no proof of stability whatsoever for it. Better try it with uncritical data partitions (portage tree, distfiles) first. Yet if you do, the latest GRUB boots from ext4 partitions, so you don't need to separate /boot. I heard it can boot from ext4 using a patch that is not 100% safe... Does anyone have positive (or negative) experience with that? --- TopperH http://topperh.blogspot.com signature.asc Description: PGP signature
Re: [gentoo-user] Resizing physical volume for lvm.
On Sun, 22 Mar 2009 22:35:35 +0200 Alan McKinnon alan.mckin...@gmail.com wrote: On Sunday 22 March 2009 22:15:14 Momesso Andrea wrote: My current setup is: Device Boot Start End Blocks Id System /dev/sda1 * 1289423246023+ 83 Linux /dev/sda228953381 3911827+ 82 Linux swap /Solaris /dev/sda33382 24804 172080247+ 83 Linux /dev/sda4 24805 3040144957902+ 83 Linux where sda3 is an lvm volume and sda4 is free space. I'd like to to merge sda3 and sda4 into a single partition without losing the data on it, but I'm not sure if it is possible. My guess is that I can use fdisk to delete sda4 and sda3, create a sda3 partition starting at 3382 and ending at 30401, then use pvresize to enlarge it. Correct. That's all there is to it. This is from man pvresize: Expand the PV on /dev/sda1 after enlarging the partition with fdisk: pvresize /dev/sda1 Is that going to work or I'm going to lose all the data? Your data is safe if you do exactly the steps you said above. pvresize /dev/sda3 /dev/sda3: too many metadata areas for pvresize Looks like I cannot expand it... --- TopperH http://topperh.blogspot.com signature.asc Description: PGP signature
Re: [gentoo-user] Resizing physical volume for lvm.
On Sun, 2009-03-22 at 21:58 +0100, Momesso Andrea wrote: On Sun, 22 Mar 2009 20:49:36 + Neil Bothwick n...@digimed.co.uk wrote: On Sun, 22 Mar 2009 21:39:29 +0100, Dirk Heinrichs wrote: P.S. I'm not using vgextend to simply add sda4 to the lvm because I might want to migrate my root (sda1) to ext4 Why do want to do that? ext4 is just a couple of months old and there's no proof of stability whatsoever for it. Better try it with uncritical data partitions (portage tree, distfiles) first. Yet if you do, the latest GRUB boots from ext4 partitions, so you don't need to separate /boot. I heard it can boot from ext4 using a patch that is not 100% safe... Does anyone have positive (or negative) experience with that? You don't need to apply any patch; it comes with GRUB. I didn't get any WARNING: this patch will fuck your box messages when I installed it. In my experience I haven't seen any difference between it and the 25 or so other patches that Gentoo applies to GRUB.
Re: [gentoo-user] Time to move on?
On Sun, Mar 22, 2009 at 1:45 PM, Neil Bothwick n...@digimed.co.uk wrote: On Sun, 22 Mar 2009 13:23:38 -0700, Mark Knecht wrote: 2) emerge -DuN world building lots of packages that are already on the system but the LoG has apparently changed flags so emerge wants to rebuild them. Don't use --newuse, use --reinstall changed-use. -- Neil Bothwick Neil, As always, thanks. So, is this option new in the last few years or something? I haven't seen it before. Also, it seems that there's no shortcut for that command so instead of emerge -pvDuN @world if I understand then I might try emerge -pvDu --reinstall changed-use @world ??? Certainly a lot more typing. ;-) Thanks. I'll check it out on the next round up updates. Cheers, Mark
Re: [gentoo-user] Time to move on?
On Sun, Mar 22, 2009 at 1:55 PM, Albert Hopkins mar...@letterboxes.org wrote: On Sun, 2009-03-22 at 13:23 -0700, Mark Knecht wrote: Personally I think we're in one of those unfortunate periods of time where there is a relatively high number of issues. I'm seeing it on all my machines. It's taking far more of my time to deal with this than I wish it would. 1) ntp-update problems at boot time. Hmm, I'm not having any ntp-update problems on my machines. Have you submitted a bug report or searched the bug database? Obviously this isn't happening for everyone so if the right people don't know about it then you can't expect it to get fixed. There's something going on here but I haven't tried to debug it. It's more like ntp isn't finding servers. some machines work. Others don't. timeouts waiting to boot. I need to find out where server names are set and then see if there is a difference between all my machines. No, I haven't filed a bug because: 1) I haven't figured out if it's my problem or Gentoo's yet 2) I have the impression that no one is reading or responding to bug reports these days, based on my generally negative view of how portage is being handled. But that's just my impression and not really worthy of a discussion because it's all free labor and free software to me so why should I complain? 2) emerge -DuN world building lots of packages that are already on the system but the LoG has apparently changed flags so emerge wants to rebuild them. But that's what -N does! That's what it's documented to do. RTFM. If you don't want that behavior then don't use -N. Either use --reinstall changed-use or don't use any USE-specific flags. Personally I just let portage re-install as it doesn't really change anything if you haven't changed your use flags. Thanks. 3) New and unclear (to me) messages about portage flag overrides caused by overlays I've been using for a while. These are probably warnings about overlays overriding settings in the regular portage profile. Some overlays do this. It's just a fact of life. Again, if you don't want to deal with it then don't use overlays or at least choose overlays that don't do potentially bad things. It's not the Gentoo devs responsibility if you bring in 3rd-party overlays that change stuff. So with the possible exception of #1, these appear to be but it hurts when I do that problems, not issues with Gentoo stability. I won't bother responding to the above. Thanks, Mark
Re: [gentoo-user] Time to move on?
On 22 Mar 2009, at 22:07, Mark Knecht wrote: ... 1) ntp-update problems at boot time. Hmm, I'm not having any ntp-update problems on my machines. Have you submitted a bug report or searched the bug database? Obviously this isn't happening for everyone so if the right people don't know about it then you can't expect it to get fixed. There's something going on here but I haven't tried to debug it. It's more like ntp isn't finding servers. some machines work. Others don't. timeouts waiting to boot. I need to find out where server names are set and then see if there is a difference between all my machines. Not sure if this helps: http://forum.soft32.com/linux/gentoo-user-init-ntpd-ntp-client-ftopict476010.html Stroller.
Re: [gentoo-user] Time to move on?
Neil Bothwick wrote: On Sun, 22 Mar 2009 15:02:12 -0700, Mark Knecht wrote: Also, it seems that there's no shortcut for that command so instead of emerge -pvDuN @world if I understand then I might try emerge -pvDu --reinstall changed-use @world ??? Certainly a lot more typing. ;-) Use an alias and it's less typing. Or add it to make.conf. I think that would work too. Dale :-) :-)
Re: [gentoo-user] Time to move on?
On Sun, 22 Mar 2009 19:26:40 -0500, Dale wrote: emerge -pvDu --reinstall changed-use @world ??? Certainly a lot more typing. ;-) Use an alias and it's less typing. Or add it to make.conf. I think that would work too. It would work, every time you called emerge, whether you wanted that option or not.I prefer to have aliases for commands with options that I call often, but not every time. -- Neil Bothwick Windows Error #09: Game Over. Exiting Windows. signature.asc Description: PGP signature
Re: [gentoo-user] Time to move on?
Neil Bothwick wrote: On Sun, 22 Mar 2009 19:26:40 -0500, Dale wrote: emerge -pvDu --reinstall changed-use @world ??? Certainly a lot more typing. ;-) Use an alias and it's less typing. Or add it to make.conf. I think that would work too. It would work, every time you called emerge, whether you wanted that option or not.I prefer to have aliases for commands with options that I call often, but not every time. I'm not real familiar with aliases but know what it is. If you use the alias method, how would you disable it for a one time run? Dale :-) :-)
Re: [gentoo-user] Time to move on?
On Sun, 2009-03-22 at 22:18 -0500, Dale wrote: Neil Bothwick wrote: On Sun, 22 Mar 2009 19:26:40 -0500, Dale wrote: emerge -pvDu --reinstall changed-use @world ??? Certainly a lot more typing. ;-) Use an alias and it's less typing. Or add it to make.conf. I think that would work too. It would work, every time you called emerge, whether you wanted that option or not.I prefer to have aliases for commands with options that I call often, but not every time. I'm not real familiar with aliases but know what it is. If you use the alias method, how would you disable it for a one time run? Uh.. you don't disable it. You simply don't use the alias.
Re: [gentoo-user] start X at startup without a login manager
I think we've got too far for it, the startx in local.start does have some drawbacks, the system will become nonresponsive if I switch back to the console, and the X seems running on VT2 instead of VT7, I studied the xdm script and found I've missed lots of important steps, for stability, I tried a login manager slim that they recommended, it worked perfect for me, everything are back to normal now! thanks again guys. fei
