Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Mon, 17 Dec 2012 00:26:13 + Kevin Chadwick wrote: > On Sun, 16 Dec 2012 22:32:24 +0200 > nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > > > My thanks, too! There's nothing like reading on some actual > > experience with this. So this was once the reason to keep / > > separate. Not that important anymore (but this is still no excuse > > to force people to keep /usr in the same filesystem). > > Sorry but real world data is important and I am fully aware of the > academic theorist problems compared to practical experience but this > simply doesn't apply here. I didn't see any evidence or > argument that a larger root conducting millions more writes is as safe > as a smaller read only one perhaos not touched for months. > > The testing criteria were very generally put and just because an > earthquake hasn't hit 200 building in the last 50 years is no reason > to remove shock absorbers or other measures from sky scrapers. > I thought I was clear in that - I was my survey of my machines for my purposes only, not a formal study in any way. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Sun, 16 Dec 2012 22:32:24 +0200 nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > > Thanks for sharing your experience, and not just your emotions. One > > of my favorite quotes is, "A man with an experience is not subject > > to a man with an argument." > > My thanks, too! There's nothing like reading on some actual experience > with this. So this was once the reason to keep / separate. Not that > important anymore (but this is still no excuse to force people to keep > /usr in the same filesystem). > A rescue/maintainer mode was *very* useful in days gone by. Lately I feel that this is better served with properly built tool (SysrescueCD and friends) can be stored on disk and launched from grub. Besides, these tools do the job so much better than just about anything you would put in a production /. Just keep your tools up to date, nothing worse than finding your brand new shiny LVM is not mountable as the metadata format is too new :-) -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Sun, 16 Dec 2012 12:49:53 -0600 Bruce Hill wrote: > On Sun, Dec 16, 2012 at 05:10:43PM +0200, Alan McKinnon wrote: > > > > That was the original reason for having / and /usr separate, and it > > dates back to the early 70s. The other reason that stems from that > > time period is the size of disks we had back then - they were tiny > > and often a minimal / was all that could really fit on the primary > > system drive. > > > > Gradually over time this setup became the norm and people started to > > depend on it, and more importantly, started to believe it was > > important to retain it. It's their right to believe that. > > > > Recently I decided to measure if I still needed a separate /usr (I > > was a long time advocate of retaining it). I'm in the lucky > > position of having ~200 Linux machines, all distinctly different, > > at my disposal, so I trawled through memory and incident logs > > looking for cases where a separate /usr was crucial to recovery > > after any form of error. To my surprise, I found none at all and > > those logs go back 5 years. > > > > So I got to change my mind (not something I do very often I admit) > > and concluded that separate base and user systems (/ and /usr) was > > no longer something I needed to do - the "system" - disks, hardware > > and the software on the disks - was very reliable, and what I > > really needed was ability to boot from USB rescue disks. I did > > find, not unsurprisingly, that I also really needed /usr/local on a > > separate partition but that's because of how we install our > > in-house software here, plus our backup policies. > > > > It also goes without saying that these days we > > need /home, /var, /var/log and /tmp to all be on their own > > filesystem, and we need that more than ever. > > > > I thought I should just toss that in the ring for people who are > > undecided where they stand on the debate of separate / vs /usr. It's > > what I found on our production, dev and staging servers, plus a > > whole lot of people's personal workstations (sysadmins and devs). > > The environment is a large corporate ISP that defies > > categorization, we almost have at least one of every imaginable > > use-case for running on Linux except something in the Top 100 > > SuperComputer list. I reckon it's about as representative as I'm > > ever gonna see. > > > > People are free to draw their own conclusions as always, and real > > data is valuable in arriving at those conclusions. YMMV. > > Thanks for sharing your experience, and not just your emotions. One > of my favorite quotes is, "A man with an experience is not subject to > a man with an argument." There's a few things I completely left out - /usr/portage and /usr/distfiles - I forgot all about those. For years now I manually move those to /var as I consider /usr to be mostly read-only, plus the portage tree and distfiles are hungry. They form two cases where separate mounts are highly desirable. The other thing I didn't comment on is /usr mounted ro over NFS. The only current valid case I've heard of is school and university labs, and one of those is the only one I've ever seen. Not something I ever work with to be honest. I would like to know how prevalent /usr as an NFS mount is in the world out there. -- Alan McKinnon alan.mckin...@gmail.com
[gentoo-user] {OT} open-source: chat, tasks, resources, code
When I need a new web-based software tool, I consider writing it myself and if that isn't feasible I try to use something open-source and self-hosted. I need something for chat, task management, resource management, and code management, all for groups. I'm considering Campfire, Trello, Float, and GitHub respectively, but I thought I'd check with you guys to see if any of this is available in an open-source and self-hosted form, especially in portage. - Grant
Re: [gentoo-user] Dual or Quad CPU complications?
> On Sun, Dec 16, 2012 at 10:39 AM, J. Roeleveld wrote: >> On Sunday, December 16, 2012 01:52:46 PM Volker Armin Hemmann wrote: >>> Am Samstag, 15. Dezember 2012, 20:57:24 schrieb J. Roeleveld: >>> > Even on a system with only 2 sockets, it can be useful to have NUMA >>> > available. >>> >>> or not, because it costs you performance. >> >> When does it cost performance? >> In all situations? > > It adds some additional logic to memory allocation (put an allocation > near the process that uses it) and to process scheduling (keep the > process near its memory, but bump it to a more distant idle core if > necessary). That's the way it's supposed to work, yes :) > In all honestly, it's not a performance loss you're likely to notice, > unless you're so in need of squeezing out every spare cycle that you > most definitely _have_ hardware where there are disconnected memory > banks. I'm not convinced it's even measurable for us mundanes and our > hardware. I don't think I would notice it either, but as the system I have supports it, I want to use it. And then I want to be certain it actually supports it correctly. The system I'm talking about is used for testing purposes. Running multiple VMs. As far as I know, Xen has support for it, just need to configure it properly. And for this usecase, I think NUMA with only 2 physical CPUs should make a positive difference. -- Joost
Re: [gentoo-user] Re: Anyone switched to eudev yet?
Michael Orlitzky wrote: > On 12/17/2012 01:09 AM, Dale wrote: >> So, since I have /usr separate from the rest, I could mount it read only >> and reduce the chance of corruption if say my UPS failed? I already do >> this for /boot. Interesting. Very interesting indeed. >> >> If the other issues happen, computers is likely the least of our >> problems. ;-) >> > Yes, although it's unlikely that any corruption would occur in the first > place if you're not writing to the filesystem during the crash. > > But unlikely is not equal to zero. Murphy's law. o_O Dale :-) :-) -- I am only responsible for what I said ... Not for what you understood or how you interpreted my words!
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On 12/17/2012 01:09 AM, Dale wrote: > > So, since I have /usr separate from the rest, I could mount it read only > and reduce the chance of corruption if say my UPS failed? I already do > this for /boot. Interesting. Very interesting indeed. > > If the other issues happen, computers is likely the least of our > problems. ;-) > Yes, although it's unlikely that any corruption would occur in the first place if you're not writing to the filesystem during the crash.
Re: [gentoo-user] Re: Anyone switched to eudev yet?
Michael Orlitzky wrote: > On 12/17/2012 12:44 AM, Dale wrote: >> >> Question. A file system, /usr for example, is mounted read only. The >> system crashes for whatever reason such as a power failure. Since it is >> mounted read only, would there be a larger or smaller risk of corrupted >> data on that partition? From what I understand, the possible corruption >> is from files not being written to the drive but since it is mounted >> read only, then that removes that possibility. >> >> Just checking on a thought here. >> > Power failure? Your data is fine. > > But "whatever reason?" Think of the possibilities! > > * The Earth stops rotating and your hard drive is flung at 67,000 > miles per hour directly into the sun. > > * Today is backwards day, and your ones and zeros have been switched. > Fsck should be able to handle this, somebody file a bug. > > * The system never really existed, it was all in your imagination. > Fade to credits. > > So, since I have /usr separate from the rest, I could mount it read only and reduce the chance of corruption if say my UPS failed? I already do this for /boot. Interesting. Very interesting indeed. If the other issues happen, computers is likely the least of our problems. ;-) Dale :-) :-) -- I am only responsible for what I said ... Not for what you understood or how you interpreted my words!
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On 12/17/2012 12:44 AM, Dale wrote: >> > > > Question. A file system, /usr for example, is mounted read only. The > system crashes for whatever reason such as a power failure. Since it is > mounted read only, would there be a larger or smaller risk of corrupted > data on that partition? From what I understand, the possible corruption > is from files not being written to the drive but since it is mounted > read only, then that removes that possibility. > > Just checking on a thought here. > Power failure? Your data is fine. But "whatever reason?" Think of the possibilities! * The Earth stops rotating and your hard drive is flung at 67,000 miles per hour directly into the sun. * Today is backwards day, and your ones and zeros have been switched. Fsck should be able to handle this, somebody file a bug. * The system never really existed, it was all in your imagination. Fade to credits.
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Dec 17, 2012 7:31 AM, "Kevin Chadwick" wrote: > > On Sun, 16 Dec 2012 22:32:24 +0200 > nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > > > My thanks, too! There's nothing like reading on some actual experience > > with this. So this was once the reason to keep / separate. Not that > > important anymore (but this is still no excuse to force people to keep > > /usr in the same filesystem). > > Sorry but real world data is important and I am fully aware of the > academic theorist problems compared to practical experience but this > simply doesn't apply here. I didn't see any evidence or > argument that a larger root conducting millions more writes is as safe > as a smaller read only one perhaos not touched for months. > > The testing criteria were very generally put and just because an > earthquake hasn't hit 200 building in the last 50 years is no reason to > remove shock absorbers or other measures from sky scrapers. > This. My desire to separate / and /usr are more for minimizing possible problems with the filesystem. Yes, I can mount /usr ro, but sooner or later I have to mount it rw, and as Murphy's Law dictates, it's exactly at that moment something bad will happen. Rgds, --
Re: [gentoo-user] Re: Anyone switched to eudev yet?
Kevin Chadwick wrote: > On Sun, 16 Dec 2012 22:32:24 +0200 > nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > >> My thanks, too! There's nothing like reading on some actual experience >> with this. So this was once the reason to keep / separate. Not that >> important anymore (but this is still no excuse to force people to keep >> /usr in the same filesystem). > Sorry but real world data is important and I am fully aware of the > academic theorist problems compared to practical experience but this > simply doesn't apply here. I didn't see any evidence or > argument that a larger root conducting millions more writes is as safe > as a smaller read only one perhaos not touched for months. > > The testing criteria were very generally put and just because an > earthquake hasn't hit 200 building in the last 50 years is no reason to > remove shock absorbers or other measures from sky scrapers. > > Question. A file system, /usr for example, is mounted read only. The system crashes for whatever reason such as a power failure. Since it is mounted read only, would there be a larger or smaller risk of corrupted data on that partition? From what I understand, the possible corruption is from files not being written to the drive but since it is mounted read only, then that removes that possibility. Just checking on a thought here. Dale :-) :-) -- I am only responsible for what I said ... Not for what you understood or how you interpreted my words!
Re: [gentoo-user] Re: /etc/portage/env for a whole category
On Mon, Dec 17, 2012 at 02:41:52AM +0200, Nikos Chantziaras wrote: > On 17/12/12 01:56, Frank Steinmetzger wrote: > > On Sun, Dec 16, 2012 at 11:11:46PM +0200, Nikos Chantziaras wrote: > >> If you really want "-ggdb" instead of just "-g", then use that instead, > >> though it will take more space for no real benefit, unless you actually > >> need the extra debug info for some reason. > > > > I wasn't really sure whether I really need it. The only thing I did try was > > -ggdb1. > > This is one of those things where if you're not sure whether you need it > or not, then you don't need it :-) Hehe. Well, sometimes I do encounter some weirdness and I would like to debug it (mostly, when a KDE program crashes, which some of those do just for fun from time to time). -- Gruß | Greetings | Qapla’ Please do not share anything from, with or about me with any Facebook service. Emacs is a great operating system, which only lacks a good editor. signature.asc Description: Digital signature
[gentoo-user] Re: /etc/portage/env for a whole category
On 17/12/12 01:56, Frank Steinmetzger wrote: On Sun, Dec 16, 2012 at 11:11:46PM +0200, Nikos Chantziaras wrote: If you really want "-ggdb" instead of just "-g", then use that instead, though it will take more space for no real benefit, unless you actually need the extra debug info for some reason. I wasn't really sure whether I really need it. The only thing I did try was -ggdb1. This is one of those things where if you're not sure whether you need it or not, then you don't need it :-)
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Sun, 16 Dec 2012 22:32:24 +0200 nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > My thanks, too! There's nothing like reading on some actual experience > with this. So this was once the reason to keep / separate. Not that > important anymore (but this is still no excuse to force people to keep > /usr in the same filesystem). Sorry but real world data is important and I am fully aware of the academic theorist problems compared to practical experience but this simply doesn't apply here. I didn't see any evidence or argument that a larger root conducting millions more writes is as safe as a smaller read only one perhaos not touched for months. The testing criteria were very generally put and just because an earthquake hasn't hit 200 building in the last 50 years is no reason to remove shock absorbers or other measures from sky scrapers.
Re: [gentoo-user] Re: /etc/portage/env for a whole category
On Sun, Dec 16, 2012 at 11:11:46PM +0200, Nikos Chantziaras wrote: > > I would like to include debug info into more of my system, but still not the > > whole userland. So I'd like to start with libs. But how do I tell portage do > > to it? > > I've been using portage/env before for selected packages, namely: > > > > $ cat /etc/portage/debug-build > > CFLAGS="-O2 -march=native -pipe -ggdb" > > CXXFLAGS="${CFLAGS}" > > FEATURES="${FEATURES} splitdebug" > > #USE="${USE} debug" > > > > and then placing a symlink… > (Do not set the "debug" USE flag. It's not for getting debug info.) If you look a second time, you will notice it is already commented out. :) > If you really want "-ggdb" instead of just "-g", then use that instead, > though it will take more space for no real benefit, unless you actually > need the extra debug info for some reason. I wasn't really sure whether I really need it. The only thing I did try was -ggdb1. Anyway, thanks for the info. -- Gruß | Greetings | Qapla’ Please do not share anything from, with or about me with any Facebook service. The advantage of RSS jokes is that you always know the newest one. signature.asc Description: Digital signature
[gentoo-user] Re: Anyone switched to eudev yet?
On 17/12/12 00:14, Volker Armin Hemmann wrote: Am Sonntag, 16. Dezember 2012, 23:19:46 schrieb Nikos Chantziaras: On 15/12/12 12:18, Volker Armin Hemmann wrote: Am Freitag, 14. Dezember 2012, 21:34:54 schrieb Kevin Chadwick: On Fri, 14 Dec 2012 08:53:35 -0800 Mark Knecht wrote: I guess the other question that's lurking here for me is why do you have /usr on a separate partition? [...] It should be moving in the other direction for stability reasons and busybox is no full answer. On OpenBSD which has the benefit of userland being part of it. All the critical single user binaries are in root and built statically as much as possible, maximising system reliability no matter the custom requirements or packages. until a flaw is found in one of the libs used and all those statically linked binaries are in danger. Well done! I don't see why this would only affect statically linked executables. If a bug is found in a library, all dynamically linked executables are affected as well. When the BSD packagers put out an update for the library, they'll also put updates for the static binaries that use it. I don't see any security issue here. with dynamically linked libs you can change just the lib, you can even just use some LD_PRELOAD workaround. As you said yourself - with statically linked libs you have to replace half of your system.. and until the binaries are ready for distribution you can't even work around it. Or you wait for the update by the vendor of your OS, which is what people do. Also, the few critical system binaries that are required to just get a shell and fix the system, are not "half of your system."
Re: [gentoo-user] Re: Anyone switched to eudev yet?
Am Sonntag, 16. Dezember 2012, 23:19:46 schrieb Nikos Chantziaras: > On 15/12/12 12:18, Volker Armin Hemmann wrote: > > Am Freitag, 14. Dezember 2012, 21:34:54 schrieb Kevin Chadwick: > >> On Fri, 14 Dec 2012 08:53:35 -0800 > >> > >> Mark Knecht wrote: > >>> I guess the other question that's lurking here for me is why do you > >>> have /usr on a separate partition? What's the usage model that drives > >>> a person to do that? The most I've ever done is move /usr/portage and > >>> /usr/src to other places. My /usr never has all that much in it beyond > >>> those two directories, along with maybe /usr/share. Would it not be > >>> easier for you in the long run to move /usr back to / and not have to > >>> deal with this question at all? > >> > >> It should be moving in the other direction for stability reasons and > >> busybox is no full answer. > >> > >> On OpenBSD which has the benefit of userland being part of it. All the > >> critical single user binaries are in root and built statically as much > >> as possible, maximising system reliability no matter the custom > >> requirements or packages. > > > > until a flaw is found in one of the libs used and all those statically > > linked binaries are in danger. Well done! > > I don't see why this would only affect statically linked executables. > If a bug is found in a library, all dynamically linked executables are > affected as well. When the BSD packagers put out an update for the > library, they'll also put updates for the static binaries that use it. > > I don't see any security issue here. with dynamically linked libs you can change just the lib, you can even just use some LD_PRELOAD workaround. As you said yourself - with statically linked libs you have to replace half of your system.. and until the binaries are ready for distribution you can't even work around it. -- #163933
Re: [gentoo-user] Re: /usr/share/doc/openrc/net.example not found
On Sun, 16 Dec 2012 12:50:02 -0800, Bryan Gardiner wrote: > What about software that refers to files in /usr/share/doc? I can't > remember what program it was, but one I used a while back had a Help > menu item that tried to open a manual living there, but of course the > manual was compressed and under a different name and it couldn't be > found. You can, at least with HTML documentation, have it linked to a consistent location, see DOC_SYMLINKS_DIR in man make.conf. > It's a minor inconvenience, but should files be left > uncompressed for cases like this? See PORTAGE_COMPRESS and associated variables in the same man page. -- Neil Bothwick Beware! The end is... signature.asc Description: PGP signature
[gentoo-user] Re: Anyone switched to eudev yet?
On 2012-12-16, Nikos Chantziaras wrote: > On 15/12/12 12:18, Volker Armin Hemmann wrote: >> Am Freitag, 14. Dezember 2012, 21:34:54 schrieb Kevin Chadwick: >> >>> On OpenBSD which has the benefit of userland being part of it. All the >>> critical single user binaries are in root and built statically as much >>> as possible, maximising system reliability no matter the custom >>> requirements or packages. >> >> until a flaw is found in one of the libs used and all those statically linked >> binaries are in danger. Well done! > > I don't see why this would only affect statically linked > executables. If a bug is found in a library, all dynamically linked > executables are affected as well. When the BSD packagers put out an > update for the library, they'll also put updates for the static > binaries that use it. > > I don't see any security issue here. Even more than that, if a flaw is found, no matter if those are statically or dinamically linked, the flaw exists both ways, and can be exploited in both scenarios. About replacing, you can just replace all those binaries like you would replace the dynamically linkable one. But you'd have to consider that the flaw may have been exploited in both scenarios. -- Nuno Silva (aka njsg) http://njsg.sdf-eu.org/
[gentoo-user] Re: Anyone switched to eudev yet?
On 15/12/12 12:18, Volker Armin Hemmann wrote: Am Freitag, 14. Dezember 2012, 21:34:54 schrieb Kevin Chadwick: On Fri, 14 Dec 2012 08:53:35 -0800 Mark Knecht wrote: I guess the other question that's lurking here for me is why do you have /usr on a separate partition? What's the usage model that drives a person to do that? The most I've ever done is move /usr/portage and /usr/src to other places. My /usr never has all that much in it beyond those two directories, along with maybe /usr/share. Would it not be easier for you in the long run to move /usr back to / and not have to deal with this question at all? It should be moving in the other direction for stability reasons and busybox is no full answer. On OpenBSD which has the benefit of userland being part of it. All the critical single user binaries are in root and built statically as much as possible, maximising system reliability no matter the custom requirements or packages. until a flaw is found in one of the libs used and all those statically linked binaries are in danger. Well done! I don't see why this would only affect statically linked executables. If a bug is found in a library, all dynamically linked executables are affected as well. When the BSD packagers put out an update for the library, they'll also put updates for the static binaries that use it. I don't see any security issue here.
[gentoo-user] Re: /etc/portage/env for a whole category
On 16/12/12 22:19, Frank Steinmetzger wrote: Hello again, I would like to include debug info into more of my system, but still not the whole userland. So I'd like to start with libs. But how do I tell portage do to it? I've been using portage/env before for selected packages, namely: $ cat /etc/portage/debug-build CFLAGS="-O2 -march=native -pipe -ggdb" CXXFLAGS="${CFLAGS}" FEATURES="${FEATURES} splitdebug" #USE="${USE} debug" and then placing a symlink at portage/env// to point at the file. But how can I tell portage to use it for whole categories, or even a group of categories like *libs*? First, clean up all current settings for this (delete all your current symlinks and such.) Then, create a *.conf file in /etc/portage/env/. The name can whatever you want. In this case, "splitdebug.conf" seems appropriate. So make a /etc/portage/env/splitdebug.conf" file with the following contents: CFLAGS="${CFLAGS} -g" CXXFLAGS="${CXXFLAGS} -g" FEATURES="${FEATURES} splitdebug" (Do not set the "debug" USE flag. It's not for getting debug info.) If you really want "-ggdb" instead of just "-g", then use that instead, though it will take more space for no real benefit, unless you actually need the extra debug info for some reason. Note how the file appends the "-g" flag to the existing flags. This way, your usual flags from make.conf are picked up; you don't have manually keep them in sync. Next, create the file "/etc/portage/package.env". In that file, you can list packages that should use a *.conf file you created. In this case, you can do: sys-libs/glibc splitdebug.conf media-libs/libsdl splitdebug.conf # etc. However, portage now also supports wildcards. So you can do stuff like: sys-libs/* splitdebug.conf media-libs/* splitdebug.conf or even: *-libs/* splitdebug.conf
Re: [gentoo-user] eudev
On Thu, 13 Dec 2012 15:39:59 + (UTC) James wrote: > Upon syncing, my system wants to upgrade to eudev. > > [blocks B] sys-fs/udev ("sys-fs/udev" is blocking sys-fs/eudev-0) > > > Not much out there; but I gleaned it is for those > that insist on a separate partition for /var and /usr. > Any other motivating reasons? > > equery depends eudev > * These packages depend on eudev: > virtual/udev-196 > (>=sys-fs/eudev-1_beta1[gudev?,hwdb?,introspection?,keymap? > ,selinux?,static-libs?]) > > > I really do not want eudev, at this time. I just recovered > a system that is now running sys-fs/udev-196-r1. > > I did recently put these into my package.keywords. > > =sys-fs/udev-196-r1 ~amd64 > =virtual/udev-196 ~amd64 > =sys-fs/udev-init-scripts-17-r1 ~amd64 > > But I do not want to go to eudev (not till it's sable > and necessary. > > Is this the best (most current) info on setting up udev-196 ? > http://wiki.gentoo.org/wiki/Udev > > > Some discussion and guidance would be keenly appreciated. > > > cautiously, > James Hi James, My guess is that you've unmasked sys-fs/udev-196 only partially. Portage tries to calculate the dependencies for it and finds that something is still missing (e.g. you need to ~amd64 more packages) so Portage stops with sys-fs/udev and tries to satisfy virtual/udev with eudev instead. Try an "emerge -pv =sys-fs/udev-196-r1" and see if that gives any reason why Portage isn't happy with it. HTH, Bryan
Re: [gentoo-user] Re: Anyone switched to eudev yet?
Nuno J. Silva wrote: > My thanks, too! There's nothing like reading on some actual experience > with this. So this was once the reason to keep / separate. Not that > important anymore (but this is still no excuse to force people to keep > /usr in the same filesystem). Mines on a separate partition because it is on LVM instead of a regular partition. Actually, only / and /boot is on a regular partition. Everything else is on LVM. I don't have / on LVM because I don't want to use a init thingy. I just wonder, how many people still have /usr on a separate partition. Like most things, there is no way to really know. Dale :-) :-) -- I am only responsible for what I said ... Not for what you understood or how you interpreted my words!
Re: [gentoo-user] Re: /usr/share/doc/openrc/net.example not found
On Sat, 15 Dec 2012 10:08:54 +0200 nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > On 2012-12-15, Chris Stankevitz wrote: > > > Hello, > > > > The file > > > > /etc/conf.d/net > > > > reports that I can seen an example format at this location: > > > > /usr/share/doc/openrc/net.example > > As dale found, it's under a compression suffix. In fact, most (all?) > of the stuff that goes under /usr/share/doc is compressed by default > under gentoo. This used to be gzip -5, and was then changed to bzip > -9, and you can change it to anything else, including no compression > at all. What about software that refers to files in /usr/share/doc? I can't remember what program it was, but one I used a while back had a Help menu item that tried to open a manual living there, but of course the manual was compressed and under a different name and it couldn't be found. It's a minor inconvenience, but should files be left uncompressed for cases like this? Cheers, Bryan
[gentoo-user] Re: Anyone switched to eudev yet?
On 2012-12-16, Bruce Hill wrote: > On Sun, Dec 16, 2012 at 05:10:43PM +0200, Alan McKinnon wrote: >> >> That was the original reason for having / and /usr separate, and it >> dates back to the early 70s. The other reason that stems from that time >> period is the size of disks we had back then - they were tiny and often >> a minimal / was all that could really fit on the primary system drive. >> >> Gradually over time this setup became the norm and people started to >> depend on it, and more importantly, started to believe it was important >> to retain it. It's their right to believe that. >> >> Recently I decided to measure if I still needed a separate /usr (I was >> a long time advocate of retaining it). I'm in the lucky position of >> having ~200 Linux machines, all distinctly different, at my disposal, >> so I trawled through memory and incident logs looking for cases where a >> separate /usr was crucial to recovery after any form of error. To my >> surprise, I found none at all and those logs go back 5 years. >> >> So I got to change my mind (not something I do very often I admit) and >> concluded that separate base and user systems (/ and /usr) was no >> longer something I needed to do - the "system" - disks, hardware and >> the software on the disks - was very reliable, and what I really needed >> was ability to boot from USB rescue disks. I did find, not >> unsurprisingly, that I also really needed /usr/local on a separate >> partition but that's because of how we install our in-house software >> here, plus our backup policies. >> >> It also goes without saying that these days we >> need /home, /var, /var/log and /tmp to all be on their own filesystem, >> and we need that more than ever. >> >> I thought I should just toss that in the ring for people who are >> undecided where they stand on the debate of separate / vs /usr. It's >> what I found on our production, dev and staging servers, plus a whole >> lot of people's personal workstations (sysadmins and devs). The >> environment is a large corporate ISP that defies categorization, we >> almost have at least one of every imaginable use-case for running on >> Linux except something in the Top 100 SuperComputer list. I reckon it's >> about as representative as I'm ever gonna see. >> >> People are free to draw their own conclusions as always, and real data >> is valuable in arriving at those conclusions. YMMV. > > Thanks for sharing your experience, and not just your emotions. One of my > favorite quotes is, "A man with an experience is not subject to a man with an > argument." My thanks, too! There's nothing like reading on some actual experience with this. So this was once the reason to keep / separate. Not that important anymore (but this is still no excuse to force people to keep /usr in the same filesystem). -- Nuno Silva (aka njsg) http://njsg.sdf-eu.org/
[gentoo-user] /etc/portage/env for a whole category
Hello again, I would like to include debug info into more of my system, but still not the whole userland. So I'd like to start with libs. But how do I tell portage do to it? I've been using portage/env before for selected packages, namely: $ cat /etc/portage/debug-build CFLAGS="-O2 -march=native -pipe -ggdb" CXXFLAGS="${CFLAGS}" FEATURES="${FEATURES} splitdebug" #USE="${USE} debug" and then placing a symlink at portage/env// to point at the file. But how can I tell portage to use it for whole categories, or even a group of categories like *libs*? As always, thanks for your insight. -- Gruß | Greetings | Qapla’ Please do not share anything from, with or about me with any Facebook service. “So the Americans are worried about having autistic Englishmen hacking into the Pentagon Computer, but they are seriously considering Mitt Romney having control of the Button.” – Jeremy Hardy about the extradition of the UFO hacker. signature.asc Description: Digital signature
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Sun, Dec 16, 2012 at 05:10:43PM +0200, Alan McKinnon wrote: > > That was the original reason for having / and /usr separate, and it > dates back to the early 70s. The other reason that stems from that time > period is the size of disks we had back then - they were tiny and often > a minimal / was all that could really fit on the primary system drive. > > Gradually over time this setup became the norm and people started to > depend on it, and more importantly, started to believe it was important > to retain it. It's their right to believe that. > > Recently I decided to measure if I still needed a separate /usr (I was > a long time advocate of retaining it). I'm in the lucky position of > having ~200 Linux machines, all distinctly different, at my disposal, > so I trawled through memory and incident logs looking for cases where a > separate /usr was crucial to recovery after any form of error. To my > surprise, I found none at all and those logs go back 5 years. > > So I got to change my mind (not something I do very often I admit) and > concluded that separate base and user systems (/ and /usr) was no > longer something I needed to do - the "system" - disks, hardware and > the software on the disks - was very reliable, and what I really needed > was ability to boot from USB rescue disks. I did find, not > unsurprisingly, that I also really needed /usr/local on a separate > partition but that's because of how we install our in-house software > here, plus our backup policies. > > It also goes without saying that these days we > need /home, /var, /var/log and /tmp to all be on their own filesystem, > and we need that more than ever. > > I thought I should just toss that in the ring for people who are > undecided where they stand on the debate of separate / vs /usr. It's > what I found on our production, dev and staging servers, plus a whole > lot of people's personal workstations (sysadmins and devs). The > environment is a large corporate ISP that defies categorization, we > almost have at least one of every imaginable use-case for running on > Linux except something in the Top 100 SuperComputer list. I reckon it's > about as representative as I'm ever gonna see. > > People are free to draw their own conclusions as always, and real data > is valuable in arriving at those conclusions. YMMV. Thanks for sharing your experience, and not just your emotions. One of my favorite quotes is, "A man with an experience is not subject to a man with an argument." -- Happy Penguin Computers >') 126 Fenco Drive ( \ Tupelo, MS 38801 ^^ supp...@happypenguincomputers.com 662-269-2706 662-205-6424 http://happypenguincomputers.com/ Don't top-post: http://en.wikipedia.org/wiki/Top_post#Top-posting
Re: [gentoo-user] Dual or Quad CPU complications?
On Sun, Dec 16, 2012 at 10:39 AM, J. Roeleveld wrote: > On Sunday, December 16, 2012 01:52:46 PM Volker Armin Hemmann wrote: >> Am Samstag, 15. Dezember 2012, 20:57:24 schrieb J. Roeleveld: >> > Even on a system with only 2 sockets, it can be useful to have NUMA >> > available. >> >> or not, because it costs you performance. > > When does it cost performance? > In all situations? It adds some additional logic to memory allocation (put an allocation near the process that uses it) and to process scheduling (keep the process near its memory, but bump it to a more distant idle core if necessary). In all honestly, it's not a performance loss you're likely to notice, unless you're so in need of squeezing out every spare cycle that you most definitely _have_ hardware where there are disconnected memory banks. I'm not convinced it's even measurable for us mundanes and our hardware. > >> And while the starting questions were not stupid this thread is overflowing >> with stupid answers. > > Matter of opinion... Indeed. -- :wq
Re: [gentoo-user] Dual or Quad CPU complications?
On Sunday, December 16, 2012 01:52:46 PM Volker Armin Hemmann wrote: > Am Samstag, 15. Dezember 2012, 20:57:24 schrieb J. Roeleveld: > > Even on a system with only 2 sockets, it can be useful to have NUMA > > available. > > or not, because it costs you performance. When does it cost performance? In all situations? > And while the starting questions were not stupid this thread is overflowing > with stupid answers. Matter of opinion...
Re: [gentoo-user] Re: Anyone switched to eudev yet?
On Sat, 15 Dec 2012 10:16:05 +0200 nunojsi...@ist.utl.pt (Nuno J. Silva) wrote: > On 2012-12-14, Mark Knecht wrote: > > > I guess the other question that's lurking here for me is why do you > > have /usr on a separate partition? What's the usage model that > > drives a person to do that? The most I've ever done is > > move /usr/portage and /usr/src to other places. My /usr never has > > all that much in it beyond those two directories, along with > > maybe /usr/share. Would it not be easier for you in the long run to > > move /usr back to / and not have to deal with this question at all? > > I may be wrong in this one, but the idea I have is that your regular > applications (so, most of them) all lie under /usr/ -- /lib /bin and > others are for essential system tools. > That was the original reason for having / and /usr separate, and it dates back to the early 70s. The other reason that stems from that time period is the size of disks we had back then - they were tiny and often a minimal / was all that could really fit on the primary system drive. Gradually over time this setup became the norm and people started to depend on it, and more importantly, started to believe it was important to retain it. It's their right to believe that. Recently I decided to measure if I still needed a separate /usr (I was a long time advocate of retaining it). I'm in the lucky position of having ~200 Linux machines, all distinctly different, at my disposal, so I trawled through memory and incident logs looking for cases where a separate /usr was crucial to recovery after any form of error. To my surprise, I found none at all and those logs go back 5 years. So I got to change my mind (not something I do very often I admit) and concluded that separate base and user systems (/ and /usr) was no longer something I needed to do - the "system" - disks, hardware and the software on the disks - was very reliable, and what I really needed was ability to boot from USB rescue disks. I did find, not unsurprisingly, that I also really needed /usr/local on a separate partition but that's because of how we install our in-house software here, plus our backup policies. It also goes without saying that these days we need /home, /var, /var/log and /tmp to all be on their own filesystem, and we need that more than ever. I thought I should just toss that in the ring for people who are undecided where they stand on the debate of separate / vs /usr. It's what I found on our production, dev and staging servers, plus a whole lot of people's personal workstations (sysadmins and devs). The environment is a large corporate ISP that defies categorization, we almost have at least one of every imaginable use-case for running on Linux except something in the Top 100 SuperComputer list. I reckon it's about as representative as I'm ever gonna see. People are free to draw their own conclusions as always, and real data is valuable in arriving at those conclusions. YMMV. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Dual or Quad CPU complications?
Am Samstag, 15. Dezember 2012, 20:57:24 schrieb J. Roeleveld: > On Saturday, December 15, 2012 11:46:36 AM Grant wrote: > > > You have to buy NUMA hardware. If the hardware you buys does not scream > > > > NUMA > > > > > at you, you don't have it. It is really that simple. > > > > > > Multicore, multisocket systems MIGHT be NUMA systems - but that is not a > > > guarantee. Now can this stupid thread please die away? > > > > I guess the question seems stupid if you already know the answer. > > There are no stupid questions, only stupid answers... > > Even on a system with only 2 sockets, it can be useful to have NUMA > available. or not, because it costs you performance. And while the starting questions were not stupid this thread is overflowing with stupid answers. -- #163933
Re: [gentoo-user] Anyone switched to eudev yet?
Am Samstag, 15. Dezember 2012, 17:43:05 schrieb Kevin Chadwick: > On Sat, 15 Dec 2012 11:18:25 +0100 > > Volker Armin Hemmann wrote: > > > It should be moving in the other direction for stability reasons and > > > busybox is no full answer. > > > > > > On OpenBSD which has the benefit of userland being part of it. All > > > the critical single user binaries are in root and built statically > > > as much as possible, maximising system reliability no matter the > > > custom requirements or packages. > > > > until a flaw is found in one of the libs used and all those > > statically linked binaries are in danger. Well done! > > How unlikely and is why you have test systems. wow, so how many vulnerabilities have you found with your test systems? Just a question. And how do they help mitigate the problem? Really? Having lots of test systems help you in which way if there is a root exploit in some lib that was wisely statically linked into half of your installed apps? Please explain. Without bullshit this time. Thank you very much. At least the 'no security hole in the default install' bullshit is gone. Easy to have a 'secure' default installation if it only contains ed, tar, cp, cat and a shell. -- #163933
Re: [gentoo-user] reboot: something renaming sub-directory in /var/run???
On Dec 16, 2012 2:44 AM, "Jarry" wrote: > > Hi Gentoo-users, I have strange problem: > > "Something" is renaming /var/run/teamspeak3-server into > /var/run/teamspeak3 in every reboot! Maybe it has something > to do with udev/openrc/baselayout2, I do not know. > This is what happens: > > I installed teamspeak3-server-bin. It creates (appart from > other files/dirs) /var/run/teamspeak3-server for pid-file > as it can be found in /etc/init.d/teamspeak3-server: > > start-stop-daemon --start --quiet --background \ > --pidfile "/var/run/teamspeak3-server/server.pid" > > I can start & stop server as usually and everything works > as expected, *as long as I do not restart server*. If I do, > after reboot there is no /var/run/teamspeak3-server, but > only /var/run/teamspeak3. Now when I try to start ts3-server, > it complains: > > start-stop-daemon: fopen '/var/run/teamspeak3-server/server.pid': > No such file or directory > > Now what the hell is going on? What (and why?) is renaming > /var/run/teamspeak3-server into /var/run/teamspeak3 during > every restart? > > /run is on tmpfs but I think it should be saved & restored > during restart without any change or lost, or am I wrong? > Something is apparently broken, but I do not know what... > tmpfs is *temporary* file system. Think of... RAM disks. Its content will be lost on every boot. Rgds, --
Re: [gentoo-user] Mounting floppy disks
On Tuesday, December 11, 2012 11:59:52 PM Paul Hartman wrote: > For some reason, when I mount floppy disk (standard HD 3.5" VFAT disk) > it does nothing. No error, just nothing... I have not tried this in > well over a year, but it used to work. > > The /dev/fd0 device works normally, I can access it with mtools and > use dd and even access disks in virtual machines running mswindows. > > Furthermore, if I make an image of the disk using dd and then loop > mount that image, it works! But when I try to mount my normal > /mnt/floppy fstab entry that I've had for years, or manually "mount -t > vfat /dev/fd0 /mnt/floppy" there is disk activity, but it never shows > up in /proc/mounts and is not mounted for me afterward. If I add "-v" > to mount, it tells me that it mounted, but it didn't actually... > either that or it has unmounted itself instantly before I can see it. > > Can anyone else mount floppies? Is there some floppy magic I have > forgotten over the years? Thanks for any tips! Paul, I can't remember the last time I actually mounted a floppy disk. I found " mtools " to be very usefull for dealing with floppy-disks: [I] sys-fs/mtools Available versions: 4.0.13 4.0.15 ~4.0.16 ~4.0.17 {{X}} Installed versions: 4.0.15(10:46:41 AM 08/02/2012)(-X) Homepage:http://mtools.linux.lu/ Description: utilities to access MS-DOS disks from Unix without mounting them Not sure why it's installed on my desktop as it doesn't actually have a floppydrive ;) -- Joost
Re: [gentoo-user] Question about new USE flags for xf86-video-intel
On Fri, Dec 14, 2012 at 09:56:45AM -0800, Mark Knecht wrote > On Fri, Dec 14, 2012 at 9:44 AM, Walter Dnes wrote: > > > I'll probably be going with SNA. Does anybody else have any experience > > and/or opinions? > > I have no idea about performance but I've got udev and sna enabled > at this time. > > gandalf ~ # emerge -pv xf86-video-intel > > These are the packages that would be merged, in order: > > Calculating dependencies... done! > [ebuild R] x11-drivers/xf86-video-intel-2.20.13 USE="dri sna > udev -glamor -uxa -xvmc" 0 kB > > Total: 1 package (1 reinstall), Size of downloads: 0 kB > gandalf ~ # > > Be forewarned, or please report back if you can, but in my case on the > two Intel machines I have (one local, one my 84 year old dad uses) X11 > stopped displaying anything after this update. As a quick fix I > removed the xorg.conf file completely to get a desktop back. This was > a couple of weeks ago. Removing xorg.conf worked on both machines and > I haven't gone back to figure out what happened and how to best > address it. I was running without an xorg.conf. This is my HTPC machine, hooked up to a 50" Panasonic Plasma TV. It's native 1366x768, but it kept on coming up 1280x720. I needed an xorg.conf "Modes" line to bring up 1920x1080 or 1366x768. As root, I ran "Xorg -configure" and copied xorg.conf.new to /etc/X11/xorg.conf. X works now, if I do not specify the resolution. It also works with resolution 1920x1080, as well as the original 1280x720. SubSection "Display" Viewport 0 0 Depth 24 Modes "1920x1080" EndSubSection However, X hangs at various other resolutions, notwithstanding that they're listed in both the TV's EDID and in the xrandr output. Plus I have other issues so I'll be starting a separate thread. -- Walter Dnes I don't run "desktop environments"; I run useful applications