Re: [gentoo-user] kernel BUG when unplugging usb?
Am Donnerstag 14 Februar 2008 07:12:48 schrieb Iain Buchanan: Hi all, When I unplug my usb mouse and keyboard for longer than a few seconds (not sure of the exact time, but must be more than, say, 5 seconds) I get a BUG message in dmesg. When I plug them back in, they don't work! I just tried to unload reload usb-hid, but modprobe locked up on the reload. It appears the only way to fix it is to reboot. This happens with tuxonice sources 2.6.23-r9 but never happened with 2.6.22.* or earlier. That bug is known (search with google), it's a race condition because of proper locking in the kernel in the evdev code. It's fixed in 2.6.24. Sascha signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Re: load too high
On Thursday 14 February 2008, Iain Buchanan wrote: On Thu, 2008-02-14 at 01:20 +0200, Alan McKinnon wrote: Ahem. 'scuse me: I have 5.5G for /var/tmp Wanna guess why? well, this is Gentoo, so compile X where X=any damn large enough package probably still fits :) Openoffice for example? spot on :-) It's a throwback to the days when I DID compile OOo. Then one day I got a clue and found openoffice-bin. Building from source is cool. Building OOo yourself is just cruel. -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] Linux, reiserfs and file fragmentation
Hi, I'm not wanting to start a flame or anything but I have a question, or two, on file fragmentation. I have three hard drives here. This is how they are partitioned at the moment: [EMAIL PROTECTED] / # df Filesystem 1K-blocksUsed Available Use% Mounted on /dev/hda619530340768940011840940 40% / /dev/hda11893397741 171822 5%/boot /dev/hda748825321805104 3077428 37%/usr/portage /dev/hda848825321402172 3480360 29%/home /dev/hdb17814576815171752 6297401620%/data [EMAIL PROTECTED] / # So you won't freak out, I removed the system generated file systems. I use reiserfs on everything except /boot. It has ext2. I found a program, script really, that will tell how fragmented a file system is. This is what it reports: [EMAIL PROTECTED] / # /root/fragck.pl /usr/portage/ 2.93016639526831% non contiguous files, 1.10476912422558 average fragments. [EMAIL PROTECTED] / # /root/fragck.pl /data/ 4.64526022181977% non contiguous files, 1.08609726757175 average fragments. [EMAIL PROTECTED] / # /root/fragck.pl /home/ statfs: No such file or directory Use of uninitialized value in pattern match (m//) at /root/fragck.pl line 32. Use of uninitialized value in concatenation (.) or string at /root/fragck.pl line 41. : not understand for /home/dale/.googleearth/Registry/google/gecommonsettings/User/layers/hiking\ and\ mountain\ bike\ trails\ . 6.4034151547492% non contiguous files, 2.34827463536108 average fragments. [EMAIL PROTECTED] / # The big one: [EMAIL PROTECTED] / # /root/fragck.pl / statfs: No such file or directory sh: en_1ca_4278190280: command not found sh: wm_6: command not found sh: wp_#usr#kde#3.5#share#wallpapers#here-gear.svgz_1196427186: command not found sh: blm_0: command not found sh: 01.png: command not found Use of uninitialized value in pattern match (m//) at /root/fragck.pl line 32. Use of uninitialized value in concatenation (.) or string at /root/fragck.pl line 41. : not understand for /var/tmp/kdecache-root/background/143x115_bm_0;en_1ca_4278190280;wm_6;wp_\#usr\#kde\#3.5\#share\#wallpapers\#here-gear.svgz_1196427186;blm_0;01.png . statfs: No such file or directory sh: en_1ca_4278190280: command not found sh: wm_6: command not found sh: wp_#usr#kde#3.5#share#wallpapers#here-gear.svgz_1192962680: command not found sh: blm_0: command not found sh: 01.png: command not found Use of uninitialized value in pattern match (m//) at /root/fragck.pl line 32. Use of uninitialized value in concatenation (.) or string at /root/fragck.pl line 41. : not understand for /var/tmp/kdecache-root/background/143x115_bm_0;en_1ca_4278190280;wm_6;wp_\#usr\#kde\#3.5\#share\#wallpapers\#here-gear.svgz_1192962680;blm_0;01.png . 7.03979903669298% non contiguous files, 1.22497799280025 average fragments. [EMAIL PROTECTED] / # Please pardon the error message. I use the script, I didn't write it. :/ I understand that doing the root directory is sort of a sum of all the others so it may be a little misleading to say the least. My questions; is this badly fragmented? How can I unfragment all the files and not bork something up badly? My opinion on this tho, considering this install is about 4 years old, not to bad. I've seen worse on a windoze rig shortly after a install. ;-) Thanks for any advice/info you can provide. Let me know if you need more info too. Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re[2]: [gentoo-user] Is it sefe to unmerge?
Hi Willie, Thursday, February 14, 2008, 6:14:44 AM, you wrote: On Thu, Feb 14, 2008 at 12:58:03AM +0100, Penguin Lover Henry Gebhardt squawked: So, yes, pwdb is a runtime dependency. I don't actually run pam, so can't confirm what would happen if I remove pwdb. Holy shit, what's going on? The ebuild in the portage tree is different than the one in /var/db/pkg/. Is it normal to update an ebuild but not its revision number? Here is the diff: ---snipped--- Damn, I spoke too soon. Just re-synced, and now this pops in the Changelog for pam 10 Feb 2008; Diego PettenC3B2 [EMAIL PROTECTED] pam-0.99.8.1-r1.ebuild, pam-0.99.9.0.ebuild: Remove dependency over pwdb, pam_pwdb is no more present in PAM 0.99, so the dependency was bogus. Where did you find it? Native pam Changelog or other location? So, yeah, pwdb is not a dependency anymore and people can safely remove it. W -- Somehow I feel like I needed the attention ~Poly Chan. After his noodle and beef stir-fry stirred the fire alarm and caused two fire engines and one ambulance to gather outside our dorm door. 06-09-2002 Sortir en Pantoufles: up 433 days, 2:43 -- Sergey -- gentoo-user@lists.gentoo.org mailing list
Re[2]: [gentoo-user] Is it sefe to unmerge?
Hi Willie, Thursday, February 14, 2008, 6:19:41 AM, you wrote: On Thu, Feb 14, 2008 at 01:17:35AM +0200, Penguin Lover Alan McKinnon squawked: emerge -n pwdb Actually, don't do that. Alan gave the right answer, but it seems that my portage tree was just a few days out of date. flameeyes just removed the dependency for pwdb from pam. (See my other reply for the message. I see. So this might actually explain why depclean and equery depends gave you different answers: one was reading the new entry for pam in the tree, and one was reading the old entry for (the same version... why didn't they bump the version?) pam in /var/db. And since the version was not bumped, your emerge --update world or whatever did not think to rebuild pam (and copy the new ebuild to /var/db) IMHO it's fully incorrect. Minor version must be changed at least... A weird incident at that. Best wishes, W -- One man's vacuum is another man's sewer. ~N. Milleron Sortir en Pantoufles: up 433 days, 2:45 -- Sergey -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Is it sefe to unmerge?
2008/2/14, Willie Wong [EMAIL PROTECTED]: On Thu, Feb 14, 2008 at 12:58:03AM +0100, Penguin Lover Henry Gebhardt squawked: Holy shit, what's going on? The ebuild in the portage tree is different than the one in /var/db/pkg/. Is it normal to update an ebuild but not its revision number? Here is the diff: ---snipped--- Damn, I spoke too soon. Just re-synced, and now this pops in the Changelog for pam 10 Feb 2008; Diego PettenC3B2 [EMAIL PROTECTED] pam-0.99.8.1-r1.ebuild, pam-0.99.9.0.ebuild: Remove dependency over pwdb, pam_pwdb is no more present in PAM 0.99, so the dependency was bogus. So, yeah, pwdb is not a dependency anymore and people can safely remove it. I agree. It seems that ebuilds do change quite frequently without a revision bump. I wrote a tiny script to see all changed ebuilds. I could'nt make it just one script, but had to make two: one for makeing the diff (ebuilddifff.sh), and one to search for the ebuilds and call the former script (findebuildiffs.sh). Just put them in the same directory, and run ./findebuildiffs.sh | less in case you are interested in what has changed. From what I can see, it seems most changes are quite trivial and indeed not worth a revision bump. But sometimes, I am not so sure... For instances, glibc-2.7-r1 is now using a different patchset (version 1.6 instead of 1.4). Does anyone know what the policy is on changing ebuilds like that? ~Henry findebuildiffs.sh Description: Bourne shell script ebuilddifff.sh Description: Bourne shell script
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Hello On Thu, Feb 14, 2008 at 05:06:43AM -0600, Dale wrote: My questions; is this badly fragmented? How can I unfragment all the files and not bork something up badly? My opinion on this tho, considering this install is about 4 years old, not to bad. I've seen worse on a windoze rig shortly after a install. ;-) I would guess the fragmented files are the big ones. And, with average of 2 fragments per file, it is not too much. If you have a movie with 30MB fragments, then it is no problem. Unless you hear lot of rattling noise from the HDD, you could leave it as is. And the surest way to defragment a filesystem is take everything out and put it back again. It will write the files one after another and will have no reason to split them. -- XML is like violence. If it doesn't solve your problem, use more. Michal 'vorner' Vaner -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Michal 'vorner' Vaner wrote: Hello On Thu, Feb 14, 2008 at 05:06:43AM -0600, Dale wrote: My questions; is this badly fragmented? How can I unfragment all the files and not bork something up badly? My opinion on this tho, considering this install is about 4 years old, not to bad. I've seen worse on a windoze rig shortly after a install. ;-) I would guess the fragmented files are the big ones. And, with average of 2 fragments per file, it is not too much. If you have a movie with 30MB fragments, then it is no problem. Unless you hear lot of rattling noise from the HDD, you could leave it as is. And the surest way to defragment a filesystem is take everything out and put it back again. It will write the files one after another and will have no reason to split them. So if for example I copied everything over to a different hard drive and then copied everything back, it would be defragmented then? I would think of something like this: Boot some live CD. Mount old and backup drives. Copy old drive to a backup drive using cp -av yada yada. Make a new file system on the old drive to make sure all is clean. Copy everything back over from the backup to the old drive using cp -av yada yada. I would also take the opportunity to redo a few partitions while I was able to. The biggest slow down by the way is when logging into KDE the first time. It takes a long while and that drive is just a getting it. The light just stays on while loading everything up. Your thoughts and others if needed. Dale :-) :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Is it sefe to unmerge?
On Thu, 14 Feb 2008 12:17:03 +0100, Henry Gebhardt wrote: It seems that ebuilds do change quite frequently without a revision bump. Does anyone know what the policy is on changing ebuilds like that? http://devmanual.gentoo.org/general-concepts/ebuild-revisions/index.html Ebuilds should have their -rX incremented whenever a change is made which will make a substantial difference to what gets installed by the package — by substantial, we generally mean something for which many users would want to upgrade. This is usually for bugfixes. Simple compile fixes do not warrant a revision bump; this is because they do not affect the installed package for users who already managed to compile it. Small documentation fixes are also usually not grounds for a new revision. IOW - If the ebuild installs basically the same code, there is no need to force everyone to recompile. -- Neil Bothwick Genius is 99% inspiration and 2% arithmetic signature.asc Description: PGP signature
Re: [gentoo-user] Linux, reiserfs and file fragmentation
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, my 2 cents: | So if for example I copied everything over to a different hard drive and | then copied everything back, it would be defragmented then? I think so yes, but still I would not do it as I think you will hardly notice the difference, but there is a good chance to screw things up. | I would think of something like this: | | Boot some live CD. | Mount old and backup drives. | Copy old drive to a backup drive using cp -av yada yada. Its very important to do this as root and preserve all the file permissions and symbolic links exactly as they are on the drive. In particular the backup file system must support all this. (You cannot backup to a FAT file system, etc.) the cp option -b could help, but surely you should read man cp and man mount | Make a new file system on the old drive to make sure all is clean. | Copy everything back over from the backup to the old drive using cp -av | yada yada. | | I would also take the opportunity to redo a few partitions while I was | able to. If you do so don't forget to update /etc/fstab and the configuration of the bootloader ! | The biggest slow down by the way is when logging into KDE the first | time. It takes a long while and that drive is just a getting it. The | light just stays on while loading everything up. I personally think this is not due to fragmentation. On loading KDE just preloads some big libraries (it is a big program :) and this takes some time. Furthermore the libraries are loaded with LD_BIND_NOW=true, which makes the linker resolve all the symbols when KDE starts. (KDE takes longer to load, but later the programs are loaded faster). You can google for that to learn what it means. Hope it helps a little Thomas -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.7 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHtDqdrpEWPKIUt7MRAkV6AKCLrm/tVj7KjM4ElHCqc0Zf/gRoxwCdEI1F jtPhK0NUmeRBcP5giKMb2JI= =Dg/q -END PGP SIGNATURE- -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] How install a local ebuild?
I have a ebuild for WICD, taken from bugzilla i don't know hot to install it. man portage says emerge /path/to/ebuild is depracated. I have to make my own overlay to put there custom ebuild? which is the better way to install local ebuild's? Cheers!
Re: [gentoo-user] How install a local ebuild?
On Thu, 14 Feb 2008 09:56:23 -0300, Ale [EMAIL PROTECTED] wrote: I have a ebuild for WICD, taken from bugzilla i don't know hot to install it. man portage says emerge /path/to/ebuild is depracated. I have to make my own overlay to put there custom ebuild? which is the better way to install local ebuild's? Cheers! Think this will help http://gentoo-wiki.com/HOWTO_Create_an_Updated_Ebuild -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] How to avoid NetworkManager logs info in terminals
I get many info lines in the tty1 every time i start NM, the same happend if i add NM service at boot time. I don't like all that output, with a simple Network manager starting [OK] is enough, which i see in the terminal when i manually start the service is ok. What can i do to avoid this? The start-stop daemon have the parameter --quiet I double check /etc/rc and the VERBOSE option for this kind of services is off i tried adding a /dev/null at the end of the start-stop daemon call, but didn't work. Any clues? Cheers!
Re: [gentoo-user] How install a local ebuild?
On Thu, 14 Feb 2008 09:56:23 -0300, Ale wrote: I have a ebuild for WICD, taken from bugzilla i don't know hot to install it. man portage says emerge /path/to/ebuild is depracated. I have to make my own overlay to put there custom ebuild? which is the better way to install local ebuild's? Set PORTDIR_OVERLAY in /etc/make.conf to, say, /usr/portage/local then mkdir -p /usr/portage/local/net-misc/wicd, copy the ebuild to that directory then do ebuild /usr/portage/local/net-misc/wicd/wicd-1.3.8-r1.ebuild manifest and emerge it in the usual way. -- Neil Bothwick If God had intended Man to program, we'd be born with serial I/O ports. signature.asc Description: PGP signature
[gentoo-user] multilib support for cross compiler toolchain
Hi, I am trying to build a glibc based cross compiler toolchain with multilib support. I am using the crossdev utility for building the toolchain. Kindly help me in enabling the multilib support in gcc. Also, how do I add -list-multilib= option? Regards, Suma Sharma -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] [gentoo-users] acpi fails with status 1
Hi all! I'm trying to use acpid with my gentoo laptop; it catches all events (battery, button, ac_adapter) but it can't execute the designed script, in any case. Here's the output of tail /var/log/messages Feb 14 14:04:10 spaventapasseri acpid: received event button/power PWRF 0080 0004 Feb 14 14:04:10 spaventapasseri acpid: notifying client 5981[0:100] Feb 14 14:04:10 spaventapasseri acpid: executing action /etc/acpi/button.sh button/power PWRF 0080 0004 Feb 14 14:04:10 spaventapasseri acpid: action exited with status 1 the script is called and it will simply display Power button pressed at the moment. It has execution permissions and the path is correct. Does anyone know a solution to this problem? Thanks in advance. Davide -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] How install a local ebuild?
Thanks both! : ) Cheers!. 2008/2/14, Neil Bothwick [EMAIL PROTECTED]: On Thu, 14 Feb 2008 09:56:23 -0300, Ale wrote: I have a ebuild for WICD, taken from bugzilla i don't know hot to install it. man portage says emerge /path/to/ebuild is depracated. I have to make my own overlay to put there custom ebuild? which is the better way to install local ebuild's? Set PORTDIR_OVERLAY in /etc/make.conf to, say, /usr/portage/local then mkdir -p /usr/portage/local/net-misc/wicd, copy the ebuild to that directory then do ebuild /usr/portage/local/net-misc/wicd/wicd-1.3.8-r1.ebuild manifest and emerge it in the usual way. -- Neil Bothwick If God had intended Man to program, we'd be born with serial I/O ports.
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Donnerstag, 14. Februar 2008, Dale wrote: The biggest slow down by the way is when logging into KDE the first time. It takes a long while and that drive is just a getting it. The light just stays on while loading everything up. do you use prelink? -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Switching to hardened
Willie Wong wrote Wonko: On Mon, Feb 11, 2008 at 11:24:49PM +0100, Penguin Lover Alex Schuster squawked: I emerged -e again, this time without distcc and ccache. All compiled fine, except for media-video/mplayer-1.0_rc2_p24929-r1 (vf_decimate.c:26: error: can't find a register in class `BREG' while reloading `asm') and http://bugs.gentoo.org/show_bug.cgi?id=175627 Like you found below, it can be avoided using vanilla GCC. That is why I still only have mplayer-1.0_rc1-r2, that one compiled okay. Isn't that the version with those many security holes? But then, looking at http://www.mplayerhq.hu/design7/news.html, it seems that all versions pre r25824 have some. x11-misc/xscreensaver-5.04: lockward.c:59: error: syntax error before uint8_t Not a problem with hardened. http://bugs.gentoo.org/show_bug.cgi?id=208731 Meanwhile, downgrade to 5.03, that one works. Thanks! But most annoying is that the nvidia drivers do not seem to work. First, what card and which drivers? 01:00.0 VGA compatible controller: nVidia Corporation NV15 [GeForce2 GTS/Pro] (rev a4) I have nvidia drivers version 71.86.01 running now. I also re-compiled xorg-server, with vanilla gcc, GLX is running fine again, and I am happy. I have an old card that is not supported by drivers = 1.0.9700, so ... scratch that, I didn't notice that the versioning scheme changed. http://www.gentoo.org/doc/en/nvidia-guide.xml they refused to compile telling me that this would do more harm than good with a hardened setup. I put them into packages.unmask, now they compile and the nvidia module loads, but still X has no GLX, xorg.0.log says Failed to initialize GLX extension (NVIDIA X driver not found), This really does not sound like a hardened issue... I need to upgrade my drivers to the 96.* to see if I can reproduce your problem, but with 1.0.8776 (from two years ago) I definitely do not have your problem. Maybe I'll try again with hardened then. My experience with nvidia is that that it makes LOTS of trouble. This, and VMware, often made kerned updates a real pain for me. I often got those errors before, with the desktop profile, on different machines. glxinfo segfaults. I guess I will try to re-compile all X stuff with the vanilla gcc. glxinfo segfaulting is expected. Do you have chpax/paxctl installed? No, not yet. I must admit I do not know much about hardened yet, but I want to play around with it and get some experience, so I started with preparing the setup by setting the hardened profile and switching to a hardened kernel. I have my entire system on the hardened profile (including X and nvidia [yes, despite the warnings of the hardened team about nvidia]) and no problems. My guess is that your problem with GLX lies somewhere else. That's good to hear! So I will stick with hardened. Would it be possible to make these changes permanent, that is, can I tell portage to compile specific packages with a specific compiler? /etc/portage/package.compilerflavor or something? Don't know. On the wiki there is a way to switch CFLAGS, don't know if something like that can be used to strip SSP and/or PIC flags from the hardened. I don't find this information there, I guess I did not look hard enough. But there is /etc/portage/bashrc, I can put a little script in there, stripping those flags for the given packages. No problem. Thanks again, Wonko -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Thomas Kahle wrote: Hi, my 2 cents: | So if for example I copied everything over to a different hard drive and | then copied everything back, it would be defragmented then? I think so yes, but still I would not do it as I think you will hardly notice the difference, but there is a good chance to screw things up. | I would think of something like this: | | Boot some live CD. | Mount old and backup drives. | Copy old drive to a backup drive using cp -av yada yada. Its very important to do this as root and preserve all the file permissions and symbolic links exactly as they are on the drive. In particular the backup file system must support all this. (You cannot backup to a FAT file system, etc.) the cp option -b could help, but surely you should read man cp and man mount | Make a new file system on the old drive to make sure all is clean. | Copy everything back over from the backup to the old drive using cp -av | yada yada. | | I would also take the opportunity to redo a few partitions while I was | able to. If you do so don't forget to update /etc/fstab and the configuration of the bootloader ! | The biggest slow down by the way is when logging into KDE the first | time. It takes a long while and that drive is just a getting it. The | light just stays on while loading everything up. I personally think this is not due to fragmentation. On loading KDE just preloads some big libraries (it is a big program :) and this takes some time. Furthermore the libraries are loaded with LD_BIND_NOW=true, which makes the linker resolve all the symbols when KDE starts. (KDE takes longer to load, but later the programs are loaded faster). You can google for that to learn what it means. Hope it helps a little Thomas I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Any ideas? Is there a limit to the fragmenting smallness? Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thu, 14 Feb 2008 06:01:16 -0600 Dale [EMAIL PROTECTED] wrote: Michal 'vorner' Vaner wrote: Hello On Thu, Feb 14, 2008 at 05:06:43AM -0600, Dale wrote: My questions; is this badly fragmented? How can I unfragment all the files and not bork something up badly? My opinion on this tho, considering this install is about 4 years old, not to bad. I've seen worse on a windoze rig shortly after a install. ;-) I would guess the fragmented files are the big ones. And, with average of 2 fragments per file, it is not too much. If you have a movie with 30MB fragments, then it is no problem. Unless you hear lot of rattling noise from the HDD, you could leave it as is. And the surest way to defragment a filesystem is take everything out and put it back again. It will write the files one after another and will have no reason to split them. So if for example I copied everything over to a different hard drive and then copied everything back, it would be defragmented then? I would think of something like this: Boot some live CD. Mount old and backup drives. Copy old drive to a backup drive using cp -av yada yada. Make a new file system on the old drive to make sure all is clean. Copy everything back over from the backup to the old drive using cp -av yada yada. I would also take the opportunity to redo a few partitions while I was able to. The biggest slow down by the way is when logging into KDE the first time. It takes a long while and that drive is just a getting it. The light just stays on while loading everything up. Your thoughts and others if needed. Dale :-) :-) :-) If you haven't already done this, you could try [1] for faster KDE boot. I believe it'll bring you much bigger application start-up boost than defragmenting your FS. Please, notice that I'm not saying that defragmentation is pointless. Just the opposite: I believe fragmentation leads to a perceivable (and actually measurable) performance hit. [1] http://www.gentoo.org/doc/en/prelink-howto.xml -- Best regards, Daniel -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Volker Armin Hemmann wrote: On Donnerstag, 14. Februar 2008, Dale wrote: The biggest slow down by the way is when logging into KDE the first time. It takes a long while and that drive is just a getting it. The light just stays on while loading everything up. do you use prelink? I did use it at one time. It didn't seem to speed anything up. I can't recall when I stopped using it tho. It's been a while. Think it would help now? It is usually when I first login that it takes so long. After the first time, it only takes about 5 or 6 seconds. Not real bad but I was thinking about my poor five year old drive. o_O Thanks. Dale :-) :-) :-) -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] Mailman trouble
Hi For some reason my mailman has stopped responding... probably after a upgrade but I'm not sure. I can't seem to start or stop mailman... # /etc/init.d/mailman stop * Stopping mailman ... [ !! ] # /etc/init.d/mailman start * WARNING: mailman has already been started. I tried to start mailman manually with '/usr/lib/mailman/mailmanctl start' but got this error: Site list is missing: mailman Then I ask mailman to display my lists with '/usr/lib/mailman/bin/ list_lists' which gave me this response: No matching mailing lists found which is somewhat odd as I checked that the lists was located where they have been the whole time at '/usr/local/mailman/lists/' which was the case. I have tried to kill all mailman processes but without any luck in regards to starting the service again. My best bet is that I have missed some crucial change when accepted changes to a conf file after an upgrade using dispatch-conf, but I have no idea which. Anybody had the same problem and found a solution? Worst case scenario, how do I move my existing lists to a fresh installation of mailman? -- Regards / Venlig hilsen Johannes Skov Frandsen /You live and learn. At any rate, you live. [Marvin]/ -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] HP 530 experiences
Hello all I am about to buy this laptop in near future,good price for exact what I need. Since I am gonna run Gentoo on it I want any kind of experience opinions. Any problems with it?Drivers etc .. ? Thank you -- Amar Ćosić [EMAIL PROTECTED] [EMAIL PROTECTED] +38761240095 http://www.amar.co.ba
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Thomas Kahle wrote: Hi, my 2 cents: | So if for example I copied everything over to a different hard | drive and then copied everything back, it would be defragmented | then? I think so yes, but still I would not do it as I think you will hardly notice the difference, but there is a good chance to screw things up. Yes, everything will be defragmented. In addition, it will leave gaps between files. So if a file lateron grows it will not immediately fragment. | I would think of something like this: | | Boot some live CD. | Mount old and backup drives. | Copy old drive to a backup drive using cp -av yada yada. Its very important to do this as root and preserve all the file permissions and symbolic links exactly as they are on the drive. In particular the backup file system must support all this. (You cannot backup to a FAT file system, etc.) the cp option -b could help, but surely you should read man cp and man mount The easiest way to preserve all permissions and symlinks is to use tar instead of cp. If you do so, read man tar of course. | The biggest slow down by the way is when logging into KDE the | first time. It takes a long while and that drive is just a | getting it. The light just stays on while loading everything up. I personally think this is not due to fragmentation. On loading KDE just preloads some big libraries (it is a big program :) and this takes some time. Furthermore the libraries are loaded with LD_BIND_NOW=true, which makes the linker resolve all the symbols when KDE starts. (KDE takes longer to load, but later the programs are loaded faster). You can google for that to learn what it means. There are two ways to speed up KDE load time. First, prelink everything (something like prelink -avmR). Second, you can configure kdm to preload as much of KDE as possible. So while you are still staring at your login screen or typing your user name and password, it loads as much as it can. BTW, KDE 4 starts significantly faster than 3.5. Uwe -- Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thu, Feb 14, 2008 at 07:53:57AM -0600, Penguin Lover Dale squawked: I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Is the /data partition on reiser? Did you enable tailpacking? As I don't know what the script you ran actually does, I don't know how it handles block suballocation... Tail packing is something that can conceivably introduce data fragmentation in place of internal fragmentation. W -- Pintsize: Curses! HOURS in there and I STILL don't have mutant ice powers! Pintsize: Sorry waffles, you can't be my sidekick until I have some superhero powers to fight crime with... What? Waffle powers? Somehow I don't see soaking up syrup or browning in a toaster getting us a lot of hot supervillain ladies. Sortir en Pantoufles: up 433 days, 12:47 -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Mailman trouble
Johannes Skov Frandsen wrote: Hi For some reason my mailman has stopped responding... probably after a upgrade but I'm not sure. I can't seem to start or stop mailman... # /etc/init.d/mailman stop * Stopping mailman ... [ !! ] # /etc/init.d/mailman start * WARNING: mailman has already been started. I tried to start mailman manually with '/usr/lib/mailman/mailmanctl start' but got this error: Site list is missing: mailman Then I ask mailman to display my lists with '/usr/lib/mailman/bin/ list_lists' which gave me this response: No matching mailing lists found which is somewhat odd as I checked that the lists was located where they have been the whole time at '/usr/local/mailman/lists/' which was the case. I have tried to kill all mailman processes but without any luck in regards to starting the service again. My best bet is that I have missed some crucial change when accepted changes to a conf file after an upgrade using dispatch-conf, but I have no idea which. Anybody had the same problem and found a solution? Worst case scenario, how do I move my existing lists to a fresh installation of mailman? I'm not sure about mailman but this is how I handle a service that is giving me fits. /etc/init.d/service-name stop. Then I use ps aux and make sure there are no living processes still there and kill them if needed. Then I do a /etc/init.d/service-name zap then do a /etc/init.d/service-name start. Cross all fingers at that point. I'm not sure if it will help or not but until some mailman guru comes along, it may be worth thinking about trying. YMMV. Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Mailman trouble
On Thu, 14 Feb 2008 15:04:59 +0100 Johannes Skov Frandsen wrote: Hi For some reason my mailman has stopped responding... probably after a upgrade but I'm not sure. I can't seem to start or stop mailman... # /etc/init.d/mailman stop * Stopping mailman ... [ !! ] # /etc/init.d/mailman start * WARNING: mailman has already been started. I tried to start mailman manually with '/usr/lib/mailman/mailmanctl start' but got this error: Site list is missing: mailman Then I ask mailman to display my lists with '/usr/lib/mailman/bin/ list_lists' which gave me this response: No matching mailing lists found which is somewhat odd as I checked that the lists was located where they have been the whole time at '/usr/local/mailman/lists/' which was the case. I have tried to kill all mailman processes but without any luck in regards to starting the service again. My best bet is that I have missed some crucial change when accepted changes to a conf file after an upgrade using dispatch-conf, but I have no idea which. Anybody had the same problem and found a solution? Worst case scenario, how do I move my existing lists to a fresh installation of mailman? A couple of thoughts: Perhaps an upgrade now has mailman looking in /usr/mailman/lists rather than /usr/local/mailman/lists I often run strace -feopen program ... to determine what files 'program' is opening (or attempting to open). Doing that can help find out where mailman is looking for its lists. Lastly, configuration files are so important that I save copies of working versions. When something breaks, at least I can compare the b0rked versions to working versions. Any version control system (rcs, csv, subversion, ...) would work for this. HTH, David -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, just one more idea that came to my mind, reiserfs uses a technique to save small files in the filesystem tree which uses less disk space then. In ext3 a 1 byte file will take up 4k, while this is not the case in reiserfs. This yields a performace hit of about 5% (people say, not that i have measured anything) If you have enough free space you can disable this, make the small files consume more space again and gains some speed improvement. Another thing you could do is disable the writing of accesstimes. Read man mount how to do this. The mount options are noatime,notail. Concerning your observation I would start looking at how the fragmentation is measured. Maybe this also depends on the filesystem implementation in the kernel. Anyway: you can not get much better 1.043.. parts per file. This means that almost every file(96,12 % in your case) is contiguous. have fun | | | I did a little test. Something fishy here. I did a test with the /data | partition. I store pictures and documents there and it was fragmented. | I cp -av to another reiserfs formatted partition then remade the file | system and copied it back using basically the same command just in | reverse. This is what I got now: | | [EMAIL PROTECTED] / # /root/fragck.pl /data/ | 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. | [EMAIL PROTECTED] / # | | That is not a lot better than it was before. It was 4.6% before. How | is that? I copied it over then ran the command right after without even | touching the files. | Any ideas? Is there a limit to the fragmenting smallness? | | Dale | | :-) :-) -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.7 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHtE6PrpEWPKIUt7MRAoAhAJ4wJ1Ygs7A75ayFCIAs+uXjW+uUbwCfdeaB rDCDg4kPoAfrKbMUZdJ/EdU= =WkET -END PGP SIGNATURE- -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] How to avoid NetworkManager logs info in terminals
Hi, Exactly same problem for me. I even compiled the networkmanger package with use flag debug disable but in vain. Thanks, flukebox On Thu, Feb 14, 2008 at 6:32 PM, Ale [EMAIL PROTECTED] wrote: I get many info lines in the tty1 every time i start NM, the same happend if i add NM service at boot time. I don't like all that output, with a simple Network manager starting [OK] is enough, which i see in the terminal when i manually start the service is ok. What can i do to avoid this? The start-stop daemon have the parameter --quiet I double check /etc/rc and the VERBOSE option for this kind of services is off i tried adding a /dev/null at the end of the start-stop daemon call, but didn't work. Any clues? Cheers!
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Dale wrote: I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Before you copy back, you have to clean the old partition - either by deleting everything or by partioning it. Uwe -- Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] [gentoo-users] acpi fails with status 1
Pupino writes: I'm trying to use acpid with my gentoo laptop; it catches all events (battery, button, ac_adapter) but it can't execute the designed script, in any case. [...] the script is called and it will simply display Power button pressed at the moment. It has execution permissions and the path is correct. Um, where will it display the output? An echo would go nowhere, I think, there is no terminal associated with that process. Have tou tried something like touch /tmp/button-pressed in the script to test it is being executed? Wonko -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Mailman trouble
For some reason my mailman has stopped responding... probably after a upgrade but I'm not sure. The mailman update this week [2.1.9-r3] broke things horrendously. In short, things moved from /usr/local/mailman to /usr/mailman and /var/mailman but there wasn't a single release note about it. The solution is documented here: http://the-hug.org/opus379.html Note that the procedure documented there does not include fixing cron; you'll need to fix paths in the mailman crontab. Be very careful to note the difference between /var and /usr in the new mailman paths. Good luck, Ed -- Ed Santiago Toolsmith [EMAIL PROTECTED] -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Uwe Thiem wrote: On Thursday 14 February 2008, Dale wrote: I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Before you copy back, you have to clean the old partition - either by deleting everything or by partioning it. Uwe I just did a mkreiserfs /dev/hdb1. That should work right? Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] emerge ruby fails
I haven't had much of a chance to google this, but I did update glibc as that seems like it's related to some ruby problems. Ruby gems fails to install with the following error (this is the topmost build error?): * Messages for package dev-ruby/rubygems-0.9.4-r2: * * ERROR: dev-ruby/rubygems-0.9.4-r2 failed. * Call stack: * ebuild.sh, line 1701: Called dyn_compile thanks, Thufir -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Daniel Iliev wrote: On Thu, 14 Feb 2008 06:01:16 -0600 Dale [EMAIL PROTECTED] wrote: Michal 'vorner' Vaner wrote: Hello On Thu, Feb 14, 2008 at 05:06:43AM -0600, Dale wrote: My questions; is this badly fragmented? How can I unfragment all the files and not bork something up badly? My opinion on this tho, considering this install is about 4 years old, not to bad. I've seen worse on a windoze rig shortly after a install. ;-) I would guess the fragmented files are the big ones. And, with average of 2 fragments per file, it is not too much. If you have a movie with 30MB fragments, then it is no problem. Unless you hear lot of rattling noise from the HDD, you could leave it as is. And the surest way to defragment a filesystem is take everything out and put it back again. It will write the files one after another and will have no reason to split them. So if for example I copied everything over to a different hard drive and then copied everything back, it would be defragmented then? I would think of something like this: Boot some live CD. Mount old and backup drives. Copy old drive to a backup drive using cp -av yada yada. Make a new file system on the old drive to make sure all is clean. Copy everything back over from the backup to the old drive using cp -av yada yada. I would also take the opportunity to redo a few partitions while I was able to. The biggest slow down by the way is when logging into KDE the first time. It takes a long while and that drive is just a getting it. The light just stays on while loading everything up. Your thoughts and others if needed. Dale :-) :-) :-) If you haven't already done this, you could try [1] for faster KDE boot. I believe it'll bring you much bigger application start-up boost than defragmenting your FS. Please, notice that I'm not saying that defragmentation is pointless. Just the opposite: I believe fragmentation leads to a perceivable (and actually measurable) performance hit. [1] http://www.gentoo.org/doc/en/prelink-howto.xml Now I remember why I stopped using prelink: The only maintenance required is re-running prelink every time a library is upgraded for a pre-linked executable. I knew there was a reason I stopped. I never could remember to run it after I finished emerging stuff. It was a thought tho. Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Dale wrote: Uwe Thiem wrote: On Thursday 14 February 2008, Dale wrote: I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Before you copy back, you have to clean the old partition - either by deleting everything or by partioning it. Uwe I just did a mkreiserfs /dev/hdb1. That should work right? Actually, I meant by formatting it instead of partioning. So yes, that should work. Maybe fragch.pl is simply buggy. Uwe -- Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Dale wrote: Now I remember why I stopped using prelink: The only maintenance required is re-running prelink every time a library is upgraded for a pre-linked executable. I only prelink after major updates. Never had any problems in between. I knew there was a reason I stopped. I never could remember to run it after I finished emerging stuff. It was a thought tho. Leave the a option ou, and it prelinks only stuff that needs prelinking. You can force automatic prelinking in /etc/conf.d/prelink. Uwe Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] emerge ruby fails
Thufir wrote: I haven't had much of a chance to google this, but I did update glibc as that seems like it's related to some ruby problems. Ruby gems fails to install with the following error (this is the topmost build error?): * Messages for package dev-ruby/rubygems-0.9.4-r2: * * ERROR: dev-ruby/rubygems-0.9.4-r2 failed. * Call stack: * ebuild.sh, line 1701: Called dyn_compile thanks, Thufir It's not the topmost build error, above that. -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] slightly OT, laptop for gentoo
Hey all, I need a laptop. But the requirements are slightly odd. This is for a machine to stay in our colo cage for use as a barcode scanner, serial interface, basic GUI, and ssh server management console, etc. It needs to have USB for an eToken, PS2 for a barcode scanner, serial to manage PDUs KVM etc, on-board ethernet, and not be flimsy, big, or expensive. Obviously we can easily get a USB to serial adapter, meaning we'd need at least 2 USB ports, but a PS2 keyboard port seems a rarity now-a-days. Battery life isn't really important. Doesn't need to be a properly ruggadized, just sturdy. 15 or so screen, with a decent resolution to fit webpages and OO documents etc, 17 is too big. 512-1G of ram, so-so CPU (this P4M 1.8 I'm using is way more than powerful enough). Can anyone recommend a laptop that does all these things, runs Linux happily (Gentoo of course), and isn't a Thinkpad (I was forced to use one a few years ago, and hated it). Thanks -- Mike Williams -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] [gentoo-users] acpi fails with status 1
2008/2/14, Alex Schuster [EMAIL PROTECTED]: Pupino writes: I'm trying to use acpid with my gentoo laptop; it catches all events (battery, button, ac_adapter) but it can't execute the designed script, in any case. [...] the script is called and it will simply display Power button pressed at the moment. It has execution permissions and the path is correct. Um, where will it display the output? An echo would go nowhere, I think, there is no terminal associated with that process. Have tou tried something like touch /tmp/button-pressed in the script to test it is being executed? Wonko -- gentoo-user@lists.gentoo.org mailing list Hi Wonko, you're right. echo simply does nothing. However I'm not able to execute everything, specially I would like to inform users with something like zenity but it doesn't work... but I've found an interesting thread on the forum that might help. Thanks for your help! Davide -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Dale wrote: I did a little test. Something fishy here. I did a test with the /data partition. I store pictures and documents there and it was fragmented. I cp -av to another reiserfs formatted partition then remade the file system and copied it back using basically the same command just in reverse. This is what I got now: [EMAIL PROTECTED] / # /root/fragck.pl /data/ 3.88457269700333% non contiguous files, 1.04344379261138 average fragments. [EMAIL PROTECTED] / # That is not a lot better than it was before. It was 4.6% before. How is that? I copied it over then ran the command right after without even touching the files. Any ideas? Is there a limit to the fragmenting smallness? Don't worry about fragmentation on reiserfs. This is not a valid concept for reiser or for ext2/3. Fragmentation is problematic on Windows machines because that code is brain dead. Some people seem to assume that it must therefore be problematic on all file systems. Reiserfs is not brain dead, it is intelligent and will balance itself out over time. It also has tail packing which can make fragmentation stats look odd if enabled. Short answer: Don't worry about it. Let reiserfs do what it wants to do when it wants to do it - it is much much much better at these decisions than you will ever be ;-) -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Donnerstag, 14. Februar 2008, Dale wrote: Now I remember why I stopped using prelink: The only maintenance required is re-running prelink every time a library is upgraded for a pre-linked executable. I knew there was a reason I stopped. I never could remember to run it after I finished emerging stuff. prelink runs by cron - not after every emerge. That is reallye enough. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Uwe Thiem wrote: Yes, everything will be defragmented. In addition, it will leave gaps between files. So if a file lateron grows it will not immediately fragment. Which will cause a stupid script to report fragmentation if the author does not understand file system structure... -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Volker Armin Hemmann wrote: On Donnerstag, 14. Februar 2008, Dale wrote: Now I remember why I stopped using prelink: The only maintenance required is re-running prelink every time a library is upgraded for a pre-linked executable. I knew there was a reason I stopped. I never could remember to run it after I finished emerging stuff. prelink runs by cron - not after every emerge. That is reallye enough. May have to check this out again then, now that a cron can remember to prelink again. LOL Dang I'm getting old. :-( Dale :-) :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Alan McKinnon wrote: On Thursday 14 February 2008, Uwe Thiem wrote: Yes, everything will be defragmented. In addition, it will leave gaps between files. So if a file lateron grows it will not immediately fragment. Which will cause a stupid script to report fragmentation if the author does not understand file system structure... Yea, I have always read that Linux file systems are a lot better at taking care of fragmentation. Considering how old this install is and how much has been installed/removed/upgraded over the past several years, I think it is not to bad really. Even 10% with the number of files I have is better than fat or NTFS. My bro has XP with NTFS and it gets downright awful. For the record, I have over 502,000 files and over 49,000 directories on this system. That's less than 20,000 files that are fragmented. It's not just the OS but documents, little movies and a LOT of pictures. Maybe I just need a bigger hard drive. O_O I have two 80GB drives and a single 40GB drive. Waiting on DSL. he he he he he I also attached a copy of the program I used. I think I got it off the forums. Maybe some guru can improve it a little. ;-) Dale :-) :-) #!/usr/bin/perl -w #this script search for frag on a fs use strict; #number of files my $files = 0; #number of fragment my $fragments = 0; #number of fragmented files my $fragfiles = 0; #search fs for all file open (FILES, find . $ARGV[0] . -xdev -type f |); while (defined (my $file = FILES)) { #quote some chars in filename $file =~ s/!/\\!/g; $file =~ s/#/\\#/g; $file =~ s//\\/g; $file =~ s//\\/g; $file =~ s//\\/g; $file =~ s/\$/\\\$/g; $file =~ s/\(/\\\(/g; $file =~ s/\)/\\\)/g; $file =~ s/\|/\\\|/g; $file =~ s/'/\\'/g; $file =~ s/ /\\ /g; #nb of fragment for the file open (FRAG, filefrag $file |); my $res = FRAG; if ($res =~ m/.*:\s+(\d+) extents? found/) { my $fragment = $1; $fragments+=$fragment; if ($fragment 1) { $fragfiles++; } $files++; } else { print ($res : not understand for $file.\n); } close (FRAG); } close (FILES); print ( $fragfiles / $files * 100 . % non contiguous files, . $fragments / $files . average fragments.\n);
Re: [gentoo-user] Fake IMAP - Real IMAP
On Tuesday 12 February 2008, Grant wrote: Can you please ssh to your box and run an nmap from your box (locally)? This will answer if smtp and imap are running and if they are being filtered by your isp. I'm not sure if someone mentioned before but imap might not be configured to listen on anything besides 127.0.0.1. I wouldn't be surprised if Cox filters 25, but nmapping locally will shed some light on it. I did this and nmap reports smtp is open and no ports are filtered. So those filtered ports are all Cox-filtered I guess. You can always ring them and ask them. Or email them first. Either way read the small print in ToS and quote it to them to cut down wasted hours of communication with inept helpdesk staff. Your argument ought to be: please open port(s) 1,2,3 . . . for these IP addresses for me only, thank you. No? You can't? Can I please speak to your manager? If you articulate your requirement clearly and elevate it to a person authorised to deal with such a request you stand a better chance of succeeding. If they try to fob you off with open a business account if you want such a service, sir cut them short and say that you are not running a business, that sending your personal mail is *not* a business service and therefore they ought to redefine it in their unreasonable ToS, and that you are not an anonymous spammer but a registered user of their network. Essentially, you are asking them to circumvent a firewall security policy. Be polite but firm. BTW, blocking all and sundry from sending spam is A_Good_Thing(TM), but if they want to be more intelligent about it they should find a way of blocking all the darned owned MSWindows botnets out there that make the US No.1 in spam generated traffic. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Mailman trouble
Johannes Skov Frandsen wrote: Anybody had the same problem and found a solution? Worst case scenario, how do I move my existing lists to a fresh installation of mailman? http://forums.gentoo.org/viewtopic-t-641573-highlight-.html There are a couple twists. You'll need to update the mailman user to point to the right homedir, make sure your lists are in the right place, etc. kashani -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Is it sefe to unmerge?
2008/2/14, Neil Bothwick [EMAIL PROTECTED]: On Thu, 14 Feb 2008 12:17:03 +0100, Henry Gebhardt wrote: It seems that ebuilds do change quite frequently without a revision bump. Does anyone know what the policy is on changing ebuilds like that? http://devmanual.gentoo.org/general-concepts/ebuild-revisions/index.html -- snip -- Thanks, that cleared it up. ~Henry
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Alan McKinnon wrote: On Thursday 14 February 2008, Uwe Thiem wrote: Yes, everything will be defragmented. In addition, it will leave gaps between files. So if a file lateron grows it will not immediately fragment. Which will cause a stupid script to report fragmentation if the author does not understand file system structure... One can assume any level stupidness of writer of little perl scripts. ;-) That aside, how would gaps *between* files ever translate into fragmentation unless the author of that particular piece of software managed to kill his very last brain cell? Uwe -- Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Dale wrote: I also attached a copy of the program I used. I think I got it off the forums. Maybe some guru can improve it a little. ;-) Not me. Perl has been invented to generate reports from log files or such. It is not a general purpose language, though many sysadmins abuse it for such. I can't read my own perl scripts I have written three weeks ago. It all looks like spider legs or chicken cratches to me. Uwe -- Informal Linux Group Namibia: http://www.linux.org.na/ SysEx (Pty) Ltd.: http://www.SysEx.com.na/ -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Uwe Thiem wrote: On Thursday 14 February 2008, Dale wrote: I also attached a copy of the program I used. I think I got it off the forums. Maybe some guru can improve it a little. ;-) Not me. Perl has been invented to generate reports from log files or such. It is not a general purpose language, though many sysadmins abuse it for such. I can't read my own perl scripts I have written three weeks ago. It all looks like spider legs or chicken cratches to me. Three weeks! You remember what you coded for up to three weeks but not usually longer Wow. You are one lucky SOB. I barely manage three DAYS lately grin -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Thursday 14 February 2008, Uwe Thiem wrote: That aside, how would gaps *between* files ever translate into fragmentation unless the author of that particular piece of software managed to kill his very last brain cell? Oops. I had a brain fart there. -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] Looking for PCI-X external SATA controller
I can find a lot of cards that are almost what I want. But I have an external drive, and a PCI-X motherboard. Not internal, and not PCI-E. Anybody know of such a beast? -- Kevin O'Gorman, PhD
Re: [gentoo-user] slightly OT, laptop for gentoo
On Thu, 14 Feb 2008 17:14:58 + Mike Williams wrote: Hey all, I need a laptop. But the requirements are slightly odd. This is for a machine to stay in our colo cage for use as a barcode scanner, serial interface, basic GUI, and ssh server management console, etc. It needs to have USB for an eToken, PS2 for a barcode scanner, serial to manage PDUs KVM etc, on-board ethernet, and not be flimsy, big, or expensive. Obviously we can easily get a USB to serial adapter, meaning we'd need at least 2 USB ports, but a PS2 keyboard port seems a rarity now-a-days. Battery life isn't really important. Doesn't need to be a properly ruggadized, just sturdy. 15 or so screen, with a decent resolution to fit webpages and OO documents etc, 17 is too big. 512-1G of ram, so-so CPU (this P4M 1.8 I'm using is way more than powerful enough). Can anyone recommend a laptop that does all these things, runs Linux happily (Gentoo of course), and isn't a Thinkpad (I was forced to use one a few years ago, and hated it). Have you thought about a PS/2 to USB adapter? I'm presently using a PS/2 keyboard _and_ a PS/2 mouse connected to a single USB port (with a Y adapter -- dual PS/2 inputs and USB output). Also a USB hub might work to connect multiple USB devices to a single port. HTH, David -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Re: load too high
On Thursday 14 February 2008, Alan McKinnon wrote: On Thursday 14 February 2008, Iain Buchanan wrote: On Thu, 2008-02-14 at 01:20 +0200, Alan McKinnon wrote: Ahem. 'scuse me: I have 5.5G for /var/tmp Wanna guess why? well, this is Gentoo, so compile X where X=any damn large enough package probably still fits :) Openoffice for example? spot on :-) It's a throwback to the days when I DID compile OOo. Then one day I got a clue and found openoffice-bin. Building from source is cool. Building OOo yourself is just cruel. The cruelty is actually worse: the machines that will benefit most from an OOo compile from source, are those old, low memory, asthmatic boxen, that take two days to complete the emerge! I am tempted to start cross-compiling. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Looking for PCI-X external SATA controller
Kevin O'Gorman wrote: I can find a lot of cards that are almost what I want. But I have an external drive, and a PCI-X motherboard. Not internal, and not PCI-E. Anybody know of such a beast A quick Google led to this: http://www.newegg.com/Product/Product.aspx?Item=N82E16816124003 Be lucky, Neil -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] slightly OT, laptop for gentoo
On Thursday 14 February 2008 22:58:13 David Relson wrote: Have you thought about a PS/2 to USB adapter? I'm presently using a PS/2 keyboard _and_ a PS/2 mouse connected to a single USB port (with a Y adapter -- dual PS/2 inputs and USB output). D'you know what, I didn't even realise such a thing existed! I've probably got dozens of USB to PS2 adaptors, and never imagined the opposite. A dual PS2 to USB could well do the trick, and my local Maplin have some in stock, a bit pricey but the company will pay. Thanks very much. -- Mike Williams -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Looking for PCI-X external SATA controller
On Thursday 14 February 2008 06:08:23 pm Neil Walker wrote: Kevin O'Gorman wrote: I can find a lot of cards that are almost what I want. But I have an external drive, and a PCI-X motherboard. Not internal, and not PCI-E. Anybody know of such a beast A quick Google led to this: http://www.newegg.com/Product/Product.aspx?Item=N82E16816124003 Be lucky, Neil I'm confused... what is the diff between pi-x pci-e and pci? The card that Neil pointed to is a PCI card. Is that what he wanted? -- From the Desk of: Jerome D. McBride -- gentoo-user@lists.gentoo.org mailing list
RE: [gentoo-user] slightly OT, laptop for gentoo
Obviously we can easily get a USB to serial adapter You can still get laptops with real serial ports (HP sell them). You might want to investigate whether you can manipulate the USB serial adapter to your requirements. I tried once with setserial and it didn't work - I havent had time to look into it further (potentially driver dependent??). Minicom works fine for configuring routers/unix boxes however. absydos adam # file /dev/ttyUSB0 /dev/ttyUSB0: character special (188/0) absydos adam # setserial /dev/ttyUSB0 Cannot get serial info: Invalid argument -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Looking for PCI-X external SATA controller
On Freitag, 15. Februar 2008, Jerry McBride wrote: I'm confused... what is the diff between pi-x pci-e and pci? The card that Neil pointed to is a PCI card. Is that what he wanted? pci is a parallel bus. 32bit, 33mhz pci-x is an 64bit, 66mhz enhancement of the pci bus - backwards compatible. If you are lucky. pci-e is a serial point-to-point interface. Not compatible. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] How to avoid NetworkManager logs info in terminals
On Thu, 2008-02-14 at 10:02 -0300, Ale wrote: I get many info lines in the tty1 every time i start NM, the same happend if i add NM service at boot time. I don't like all that output, with a simple Network manager starting [OK] is enough, which i see in the terminal when i manually start the service is ok. What can i do to avoid this? what's wrong with output? can you post the output verbatim? I had a look at the init script and it doesn't seem to print much. The start-stop daemon have the parameter --quiet I double check /etc/rc and the VERBOSE option for this kind of services is off i tried adding a /dev/null at the end of the start-stop daemon call, but didn't work. not quite sure what /dev/null would do. I tried this with net.eth0: sudo /etc/init.d/net.eth0 restart /dev/null and it got rid of all the output. To be sure, you could add 21 sudo /etc/init.d/net.eth0 restart /dev/null 21 Any clues? Cheers! HTH, -- Iain Buchanan iaindb at netspace dot net dot au Advertising Rule: In writing a patent-medicine advertisement, first convince the reader that he has the disease he is reading about; secondly, that it is curable. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Re: load too high
sorry to hijack the thread even further... On Thu, 2008-02-14 at 23:04 +, Mick wrote: On Thursday 14 February 2008, Alan McKinnon wrote: On Thursday 14 February 2008, Iain Buchanan wrote: On Thu, 2008-02-14 at 01:20 +0200, Alan McKinnon wrote: Ahem. 'scuse me: I have 5.5G for /var/tmp Wanna guess why? well, this is Gentoo, so compile X where X=any damn large enough package probably still fits :) Openoffice for example? spot on :-) It's a throwback to the days when I DID compile OOo. Then one day I got a clue and found openoffice-bin. Building from source is cool. Building OOo yourself is just cruel. The cruelty is actually worse: the machines that will benefit most from an OOo compile from source, are those old, low memory, asthmatic boxen, that take two days to complete the emerge! I am tempted to start cross-compiling. I've often used distcc between amd64 and x86 machines, for example, and had no problems (except that not enough is farmed out). cya, -- Iain Buchanan iaindb at netspace dot net dot au Voiceless it cries, Wingless flutters, Toothless bites, Mouthless mutters. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] slightly OT, laptop for gentoo
On Thu, 14 Feb 2008 23:29:58 + Mike Williams wrote: On Thursday 14 February 2008 22:58:13 David Relson wrote: Have you thought about a PS/2 to USB adapter? I'm presently using a PS/2 keyboard _and_ a PS/2 mouse connected to a single USB port (with a Y adapter -- dual PS/2 inputs and USB output). D'you know what, I didn't even realise such a thing existed! I've probably got dozens of USB to PS2 adaptors, and never imagined the opposite. A dual PS2 to USB could well do the trick, and my local Maplin have some in stock, a bit pricey but the company will pay. Here in Michigan they seem a bit pricey, as well. A single PS/2 to USB adapter is a few dollars but the dual PS/2 to USB Y adapter is $16.00 (or worse). I had hoped to use one with my PS/2 only KVM but the combo doesn't work. I suspect the issue is with the KVM as I have 2 different Y adapters and neither works. Sigh :- -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] slightly OT, laptop for gentoo
On Thu, 2008-02-14 at 20:21 -0500, David Relson wrote: On Thu, 14 Feb 2008 23:29:58 + Mike Williams wrote: On Thursday 14 February 2008 22:58:13 David Relson wrote: Have you thought about a PS/2 to USB adapter? I'm presently using a PS/2 keyboard _and_ a PS/2 mouse connected to a single USB port (with a Y adapter -- dual PS/2 inputs and USB output). D'you know what, I didn't even realise such a thing existed! I've probably got dozens of USB to PS2 adaptors, and never imagined the opposite. A dual PS2 to USB could well do the trick, and my local Maplin have some in stock, a bit pricey but the company will pay. Here in Michigan they seem a bit pricey, as well. A single PS/2 to USB adapter is a few dollars but the dual PS/2 to USB Y adapter is $16.00 (or worse). I had hoped to use one with my PS/2 only KVM but the combo doesn't work. I suspect the issue is with the KVM as I have 2 different Y adapters and neither works. Sigh :- The Dell business models (precision) docking stations have 2 PS2 ports. http://cgi.ebay.com.au/DELL-Dock-Station-PR01X-LATITUDE-INSPIRON-FREE-DELIVERY_W0QQitemZ320215888408QQihZ011QQcategoryZ3709QQssPageNameZWDVWQQrdZ1QQcmdZViewItem A docking station would be an excellent idea if you need the peripherals to work without fussing with cables. If/when you do need to pick it up to move / replace something, you don't need to worry about unplugging replugging 6 different cables in the right spots... You can even tie / screw the docking station down (or get one of those D-View laptop stands) http://cgi.ebay.com.au/Dell-D-View-Laptop-Stand-NEW_W0QQitemZ200199102199QQihZ010QQcategoryZ3708QQssPageNameZWDVWQQrdZ1QQcmdZViewItem sorry about the ebay links, but dell of course doesn't list these things in their products page... -- Iain Buchanan iaindb at netspace dot net dot au The most exciting phrase to hear in science, the one that heralds new discoveries, is not Eureka! (I found it!) but That's funny ... -- Isaac Asimov -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] OpenVPN setup
On Wed, 13 Feb 2008 08:19:48 -0800 Grant [EMAIL PROTECTED] wrote: Even if you just want to encrypt some clear-text protocol that doesn't have an encrypted equivalent, a vpn is still overkill. For that you use ssh tunneling (which is essentially the same thing as an encrypted version of a protocol). 'ssh -X' is the classic example of easily tunneling a protocol that doesn't have a native encrypted equivalent. I see what you're saying. Can tunneling through ssh be made automatic so that a cron job initiates a script that opens a tunnel between the remote server and local print server and pages are printed through the tunnel? Sure. ssh is just a process after all and in principle encapsulated whatever gets put into it. All you need is a connection that isn't firewalled out and an sshd that is listening to what is coming in. ssh will even port forward for you and can be made to transform any tcp connection to appear to come from whatever port you want. What you put inside the tunnel is up to you. If the print server won't accept what is coming in, then google will find you any number of apps that will mangle the traffic. Your statement it seems like running SSH inside a VPN is better for security than running SSH on a non-standard port is non-sensical. From a security and encryption perspective, ssh and OpenVPN are exactly the same thing - stuff wrapped in an encryption layer provided by ssl, complete with exactly the same key setup should you choose to use that route. What about having ssh, imap, smtp, cups, and possibly a non-standard https port all hidden within a VPN? Should that be considered a benefit of running a VPN? I've filed the original post somewhere else and forgot the scenario :-) Is this a setup you need to be present often or even all the time? If so, you have 5 protocols in use, and setting up tunnels could become cumbersome. You might consider that it's more effort than it's worth and a VPN that is there and JustWorks(tm) is preferable. I would call that a sensible use of a VPN :-) I don't think there's a golden rule about when using a VPN is right or wrong. It's more like do the advantages outweigh the hassle of setting it up and maintaining it?. Sometimes this answer is obvious, sometimes less so. Sometimes it's a judgement call. Thanks a lot for everyone's help. Here is a more to-the-point list of what I'd like to accomplish: 1. encrypt CUPS printouts between remote server and local print server 2. add an additional layer of security around SSH and CUPS on local firewall/print server 3. add an additional layer of security around SSH, IMAP, and non-standard port HTTPS on remote server 4. enable access to SMTP on remote server for me which is blocked by my local ISP It sounds like I have 3 choices: 1. VPN 2. SSH tunneling 3. Zebedee tunneling Would all 3 of these choices accomplish all 4 requirements? I would think SSH tunneling can't really add an additional layer around SSH. Encrypted packets, encrypted? Why not? I'd like to have something I can leave up all the time so the services are always protected and I don't have to go through an extra step to use email or print from the remote server. Can all 3 of these be left up all the time? Is there any reason not to leave this type of functionality up all the time? I don't use tunnels, but leave VPN up all the time. It sounds like VPN would be the most difficult to set up and maintain, followed by SSH tunneling, followed by Zebedee tunneling. Maybe I'm wrong though. With tunneling, would I need to set up 4 or 5 different tunnels for CUPS, IMAP, SMTP, non-standard port HTTPS, and SSH (if I'm using Zebedee)? tunnels aren't configured, but would probably have to be created at boot. vpn is, I suppose, not super easy to configure. I will send you my config files though if you want. To send me mail, mail servers need to connect to my remote server's SMTP right? Would setting up a tunnel or VPN for my SMTP access interfere with that? Not if you tunnel through to the right ports - or in the case of a VPN, no. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] strange ethernet behavior with Superjmicro mb and Gentoo
On Wed, 13 Feb 2008 09:57:07 -0500 John covici [EMAIL PROTECTED] wrote: Hi. I have just gotten a computer with a Super Micro c2sbe Motherboard. Now I also bought a dual port PCI Express ethernet card. Now the normal kernel driver in my 2.6.21-gentoo-r4 does not recognize the ethernet port on the motherboard, only the dual port PCI Express card. So I found an Intel driver which seemed to be a later version of the e1000 driver and installed it. Now the strange part is that in order for the Ethernet on the mb to be recognized, I must rmmod and modprobe the module again and blacklist it from udev, although the later seemed not to do much. Does anyone know why I must do such a thing? If I don't do the rmmod and modprobe the dual port card is still recognized, but the mb one is not.. Thanks in advance for your help. You might consider asking the source of the driver. -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] Failing to build sane-backends
I haven't been able to build sane-backends. make[1]: *** No rule to make target `libsane-sane-epson2.la', needed by `all'. Stop. any thoughts? -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
Alan McKinnon wrote: On Thursday 14 February 2008, Uwe Thiem wrote: That aside, how would gaps *between* files ever translate into fragmentation unless the author of that particular piece of software managed to kill his very last brain cell? Oops. I had a brain fart there. You two are so funny. I found this too: http://www.oo-software.com/home/en/products/oodefrag/ Seems someone is trying to make money. I have also read that most Linux file systems do this automatically somehow. After doing my test, I tend to agree. So why have a commercial product for this? Is it just money? Your thoughts, humor is OK too. o_O Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Looking for PCI-X external SATA controller
Neil Walker wrote: Kevin O'Gorman wrote: I can find a lot of cards that are almost what I want. But I have an external drive, and a PCI-X motherboard. Not internal, and not PCI-E. Anybody know of such a beast A quick Google led to this: http://www.newegg.com/Product/Product.aspx?Item=N82E16816124003 Be lucky, Neil Dale makes a note of this. Questions: If I buy this card and a SATA hard drive, will I notice faster transfer speed on the drive or will the PCI bus limit it somehow? I currently get 40 to 50 MBs/sec on my IDE drives. Would this setup be any faster? Thanks Dale :-) :-) -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Linux, reiserfs and file fragmentation
On Friday 15 February 2008, Dale wrote: Alan McKinnon wrote: On Thursday 14 February 2008, Uwe Thiem wrote: That aside, how would gaps *between* files ever translate into fragmentation unless the author of that particular piece of software managed to kill his very last brain cell? Oops. I had a brain fart there. You two are so funny. Thank you. We try to please :-) I found this too: http://www.oo-software.com/home/en/products/oodefrag/ Seems someone is trying to make money. I have also read that most Linux file systems do this automatically somehow. After doing my test, I tend to agree. So why have a commercial product for this? Is it just money? Yeah, pretty much just money. Microsoft's business model is to trap the market, never perform at any level higher than mediocrity, and create an ecosystem that needs thousands of support apps just to keep the OS limping along. Then shaft all of them with vendor-lockin Coping with file fragmentation has to be one of the easiest algorithms around, it isn't even hard. Write a file, and look to see how the blocks are distributed. If it can be improved, then do so. Otherwise leave it as is But then again, if you have written a file system so that everything is just mushed onto the same device, all higeldypigeldy with no sane structure at all ... then I suppose you would need stuff like defrag to come along once a week and save your ass :-) -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] ooffice draw can't export to eps anymore
Hi all, I used to make drawings in openoffice draw, and then choose File export. From there I could choose to export them as eps, which is great for putting in my latex documents. However, that option isn't there anymore. The last timestamp on the last eps drawing I exported is May 15 2006. According to genlop that puts the OOo version at openoffice-bin-2.0.2. I'm now using openoffice-bin-2.3.1 Google seems to think I can still do it. Any ideas? thanks, -- Iain Buchanan iaindb at netspace dot net dot au Be wary of strong drink. It can make you shoot at tax collectors and miss. -- Lazarus Long, Time Enough for Love -- gentoo-user@lists.gentoo.org mailing list