[gentoo-user] net-fs/samba 4.5.16 compile fail
Hi, Discovered an issue when trying to compile net-fs/samba-4.5.16. Looks like it involves the WAF package configuration process and/or rpcgen. Problem is with the source files trying to "#include " (sm_notify.c is the first the fail), the location of the glibc SunRPC headers when it doesn't exist. Those headers actually live in "" but for whatever reason the build wasn't including it, or including "-I /usr/include/tiprc" during the build. Curiously, WAF shows that it didn't find the rpc headers (at least form my quick perusal through the build logs), but the build is still doing the "#include". Quick fix is: $ CFLAGS "-I/usr/include/tirpc -Wl,-ltirpc" emerge -1v samba BTW, no thanks to the a particular scold and martinet in #gentoo named "kurly" who seemed to be itching to kick/silence for asking simple questions (which he of course did). People like that shouldn't have ops, IMHO. - Jim
Re: [gentoo-user] Systemd upower
On 6/3/2014 16:13, Canek Peláez Valdés wrote: On Tue, Jun 3, 2014 at 5:06 PM, Alon Bar-Lev alo...@gentoo.org wrote: On Wed, Jun 4, 2014 at 12:59 AM, Canek Peláez Valdés can...@gmail.com wrote: On Tue, Jun 3, 2014 at 4:40 PM, Alan McKinnon alan.mckin...@gmail.com wrote: [...] Incidentally, what exactly is wrong with systemd writing a dhcp server client, and an ntp client? Is that project prohibited from writing such software? Are they not allowed to do it? Does it break legal laws? Is there an NDA or non-compete clause in the mix that I'm not aware of? Because they are the only things that could stop systemd from writing such code; without such prohibitions they are free to spend their time doing whatever they damn well please and if that means yet another dhcp implementation, so be it. Alan, thanks for succinctly putting why is absurd to complain about someone else's desire to write whatever code she desires to write. And to sharing it to the world! The HORROR! How *DARE* they to release their code? For free! Once again, you do not understand the claim. It is you who does not understand how software workds. See Alan response. If a user of Gentoo chooses to use non systemd profile, it means that we need to make sure systemd will not be a valid option, ever. Again, you don't understand how software works: this has nothing to do with profiles, it has to do with the fact that UPower now relies on systemd, and therefore people who has UPower installed now, *by default*, require systemd. If they don't want systemd, there is a way to do it, but it requires manual intervention since they now need to first uninstall UPower. In this case, if it is to disable the upower USE flag, or to provide alternative, block newer version, whatever make it possible to have a system working without systemd. It is provided: emerge -C upower emerge -1v upower-pm-utils It has to be done manually, though; otherwise you step on systemd users. systemd should not be visible at any time, nor its implications. Nobody is here to deal with other people's OCD. Regards. FWIW, on my system, I had to mask sys-apps/gentoo-systemd-integration for it to merge the udev update w/o trying to pull in systemd, et al. i didn't deep dive on what was trying to pull that in, but masking it (plus a ton of other stuff I have masked) prevented portage from trying to build a systemd based system.
Re: [gentoo-user] How's the openrc update going for everyone?
On 5/10/2011 18:25, Indi wrote: On Wed, May 11, 2011 at 12:20:02AM +0200, Paul Hartman wrote: On Tue, May 10, 2011 at 4:55 PM, Dale rdalek1...@gmail.com wrote: Hi folks, I was curious, what's the results of the openrc update for people that have done theirs? Is it pretty simple and just works or are there issues? I'm mostly interested in x86 and amd64 since that is what I have. Just a simple works here and I'm X86 or amd64 would be nice. List issues if you had any. I'm using ~amd64 and upgraded long, long, long ago. No problems at all during or after the upgrade. I would expect it to be even smoother process now than it was then. IIRC the biggest deal with the baselayout/openrc upgrade was that you must update a bunch of config files, which are not necessarily all blind/trivial updates. Failing to update them could make rebooting a sad experience. Same here, on x86 and ppc. Most of it was handled automatically and the rest via dispatch-conf. Works just fine. Went pretty smoothly for me following the upgrade guide. I have a gentoo based iptables firewall with a fairly complicated network setup with postup() functions. I made the mistake of taking out the BASH syntax (the surrounding parens, etc) on my postup() function based on the guide, wondering if it'd work or not, and sure enough it wanted the old BASH style syntax for those functions, but the new style (w/o parens, and quoted blocks with CRs) on the normal sections. They should probably make a note of this in the config guide. It's good to see they added in support for iproute2 rules natively instead of requiring a postup() function. I'd like to see them add a similar functionality for adding static ARP entries too (right now using my own postup() for that). -Jim smime.p7s Description: S/MIME Cryptographic Signature
Re: [gentoo-user] Re: LVM : pros cons
Dale wrote: Philip Webb wrote: 071008 Alan McKinnon wrote: This question Is LVM a good idea? keeps cropping up on mailing lists. I find this a bit strange as I find huge benefits and have yet to find a valid downside for general use. If you haven't used it, it looks like a questionable extra complexity, which could bite your fingers unexpectedly for little real gain. However, I am grateful for all the replies may decide to use it, the comments being generally reassuring as to its stability. There is a Gentoo doc re how to fit it into the install process. That's me too. I checked into it but just have not got the nerve up to switch. I have had to redo my partitions a couple times though so I could use it for sure. Maybe one of these days. Dale :-) :-) :-) I must be one of the 'crazy' ones. I run LVM2 for / under linux RAID. Even /boot is on a RAID1. I use genkernel for the proper initramfs. You just need to use dolvm2 and lvmraid=/dev/mdX ... on your boot line. smime.p7s Description: S/MIME Cryptographic Signature
Re: [gentoo-user] {OT} Webmail in portage without PHP?
kashani wrote: Grant wrote: Is there a decent webmail package in portage (or a layman overlay) that doesn't depend on PHP? what is the problem with php? Every webmail is going to need to depend on some kind of server side scripting. Nangus I am using perl and I'd rather not install and maintain PHP just so I can use Squirrelmail. It sounds like I may need to though. Does anyone prefer another webmail client to Squirrelmail? - Grant I've run squirrel, horde, and roundcube. I like roundcube best. It's probably the simplest to setup and the cleanest interface. I've set up squirrel and Horde IMP to check them out. Squirrel was relatively easy to set up compared to Horde, and via plugins, has pretty much every feature I want. Horde was a pain to set up, is big and slow, but looks a bit nicer than squirrel (haven't experimented with any skins on squirrel). I also installed ingo and turba on top of horde too, and I'm a bit disappointed. Unless there's a misconfiguration, turba doesn't seem to allow you to simply add a contact from an email message in Imp. And turba's mail filters seem fairly limited. I was expecting something more on the lines of Tbird's mail filter capabilities. I suppose you could get that if you ran procmail or maildrop, or perhaps seive (which I know nothing about, and apparently requires the cyrus imap server). Roundcube looks very promising. Looks like the project is going for a web based T-bird style client. :-) -- gentoo-user@gentoo.org mailing list
[gentoo-user] Console vanishes after X server is run
Hi, I'm having an issue with text console of one of my gentoo boxes not coming back after the X server exits. The console is there and works fine up to the point when the X server starts (for instance, when gdm comes up). After this, using the key sequences to call up the text vtys, shutting down gdm, etc, results in a black screen. I've tried the various key combos, etc, and the screen just stays black until the system is rebooted (it stays black until the system warm boots and the BIOS screen pops up), or until an X server is started up again (gdm will start up and display the login screen normally). It's as if the whatever is responsible for switching the video mode and/or FB address back to the console settings is not doing its job. I run gensplash and the vesa-tng frame buffer drivers. The graphics card I have is an old Nvidia GeForce 3, and I run the Nvidia binary drivers (x11-drivers/nvidia-drivers) under Xorg 7.0. Here is my boot line: kernel (hd0,0)/kernel-genkernel-x86-2.6.16-gentoo-r12 root=/dev/ram0 init=/linuxrc real_root=/dev/vg00/root dolvm2 lvmraid=/dev/md0 lvmraid=/dev/md1 video=vesafb:[EMAIL PROTECTED] splash=verbose,fadein,mtrr,theme:livecd-2006.0 quiet CONSOLE=/dev/tty1 initrd (hd0,0)/initramfs-genkernel-x86-2.6.16-gentoo-r12 (1280x1024-16 mode is listed in the /proc/fb0/modes file, so it's a valid BIOS mode) Any ideas ? Suggestions ? - Jim smime.p7s Description: S/MIME Cryptographic Signature
Re: [gentoo-user] booting with an lvm2 root. How do I make a working initrd?
Sascha Lucas wrote: On Fri, 23 Jun 2006, Javier Ubillos wrote: I don't want to use genkernel, but I still want it to work. you realy should use genkernel it's much easier. mkinitrd --preload md_mod --with=md_mod initrd-2.6.16-gentoo-r9 well you should not use mkinitrd, if you refuse genkernel :-). btw: md-mod is RAID (multiple devs), LVM is dm-mod (dev mapper). In conclusion: how do I get/make a working initrd so I can boot with my lvm2 root? that's much work to do it your self start building ext2 images or cpio's with static busybox lvm2 and write your own linuxrc (here you may cheat by looking into linuxrc from genkernel :-)). Don't forget to create dev-files and to include some modules. Sascha. Yes. What Sascha said. I've had many a user (in #gentoo IRC) sneer at me for using Genkernel to produce a kernel + initramfs archive to boot up my RAID + LVM2 based system. But it's far simpler than reinventing the wheel by creating all that is needed to bootstrap the RAID/LVM2 setup (static bins, linuxrc script to start everything up, create device nodes, etc, etc). I don't see much value in doing it myself when Genkernel already does it for me. Also, you can do all the kernel customization you wish (within reason) even if you use Genkernel by using one of the --menu|x|g|config flags, and/or --oldconfig. -- +---+ | Jim Burwell - Sr. Systems/Network/Security Engineer, JSBC | +---+ | I never let my schooling get in the way of my education. - Mark Twain | | UNIX was never designed to keep people from doing stupid things, because | | that policy would also keep them from doing clever things. - Doug Gwyn | | Cool is only three letters away from Fool - Mike Muir, Suicyco | | ..Government in its best state is but a necessary evil; in its worst | | state an intolerable one.. - Thomas Paine, Common Sense (1776)| +---+ | Email: [EMAIL PROTECTED] ICQ UIN: 1695089 | +---+ -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] vnc.so
Tom Smith wrote: Hi all, I installed vnc with the server USE flag enabled. I've been working on getting the vnc.so module loaded for X.org and now have it working. My question is, is this module a part of RealVNC or some other program? (I tried to locate a description of the use flags for VNC but was unable to.) Thanks in advance for the help. Yes the vnc.so X module is part of the VNC package. AFAIK its purpose is to implement the shared display # 0 functionality, where you can start a VNC session and interact with the system's local :0 X display. That is, say that you're logged into the target machine itself and have an X session running on it attached to the system's monitor/vid card/KB/mouse. The vnc.so provides the functionality to attach to THAT session. Note that not all VNC packages support this (notably TightVNC doesn't, but RealVNC seems to, although I havn't tried it. There's also 'xf4vnc' and 'x11vnc' which appear to provide the functionality). Perhaps the module is also used in the normal Xvnc session also, but TightVNC doesn't even include the module and it implements Xvnc. Normally, vncserver starts up its own X server (Xvnc) which uses a new X display number which is completely independant of any X server which may be running on the target system. Perhaps this is what you're really looking for ? If so, it's very easy. Just run something like 'vncserver -depth 24 -geometry 1024x768' and you'll be able to attach a VNC viewer to the virtual X display by connecting the viewer to hostname or IP:1. - Jim -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] Software RAID Advice Needed
[EMAIL PROTECTED] wrote: If a disk fails, your system would likely crash (due to the swap device), but would reboot in a degraded mode (no swap, slow performance, etc). You could avoid that by not using RAID for swap. Instead, use four separate swap partitions, one on each drive. As long as they all have the same priority, the kernel will share swap duties between them equally. If you make more swap partitions on more physical drives with the same priority, it is the same as swap on raid0: system strips swap across drives. And if some drive crashes and swap partition on that drive has been used, very probably system crashes too. But then reboots at least with remaining swap partitions... Yes, although there's a posibility of an 'endless crash/reboot' scenereo here, if the errors are 'soft' (e.g. not drive just vanishing). For instance, a few bad sectors develop on one of your swap partitions, the kernel can't read them, and panic/reboots. The system comes back up, the same swaps are used, and it happens again, over and over until you edit the bad partition out of the fstab. In a redundant RAID situation, I'm presuming that a bad sector or two would result in the RAID driver detaching the bad drive, and chugging along in degraded mode, where if this happened in a distrubuted swap situation, it's already 'too late', since the sectors are lost and the kernel would probably panic. There's no real benefit to using RAID for swap, unless you are limited on RAM and use swap a lot, when RAID0 may help. There is some benefit, if you use raid1 for swap. In such a case even drive failure does not cause system crash, because swap space is mirrored too. But raid1 slightly degrades swap performance... This is exactly why I'm doing RAID1 on swap. If one drive goes poof, my system stays up. Based on what this server is going to do, it should rarely use much swap, so swap performance isn't a priority for me. Plus, as you say, I believe the performance hit on swap writes (reads should actually be faster) should wind up being only a bit slower than if you were swapping to a single drive. - Jim -- +---+ | Jim Burwell - Sr. Systems/Network/Security Engineer, JSBC | +---+ | "I never let my schooling get in the way of my education." - Mark Twain | | "UNIX was never designed to keep people from doing stupid things, because | | that policy would also keep them from doing clever things." - Doug Gwyn | | "Cool is only three letters away from Fool" - Mike Muir, Suicyco | | "..Government in its best state is but a necessary evil; in its worst | | state an intolerable one.." - Thomas Paine, "Common Sense" (1776)| +---+ | Email: [EMAIL PROTECTED] ICQ UIN: 1695089 | +---+ | Reply problems ? Turn off the "sign" function in email prog. Blame MS. | +---+
Re: [gentoo-user] Software RAID Advice Needed
Richard Fish wrote: On 12/15/05, Ognjen Bezanov [EMAIL PROTECTED] wrote: I have found Linux Software RAID very useful and reliable. While probably being beaten in the performance area by hardware implementations, I just want to point out that when we are talking hardware here, we mean real hardware RAID...made by companies like 3-ware. The 'hardware' RAID in the NForce4 chipset (like just about all MB chips, and a lot of the cheap add-in cards) is just a BIOS helper...all of the actual RAID functions are expected to be implemented by the driver running on the CPU. Don't you hate how the hardware and mobo manufacturers have muddied the hardware RAID waters by marketing this sort of thing has hardware RAID (or at least implying it) ? Another thing to check out, seeing that he has a mobo with built in ghetto-RAID (TM), is dmraid. This is a device mapper implementation of RAID which makes use of various fake hardware RAID metadata to support them under Linux. Someone's also done up a Gentoo LiveCD with dmraid support on it too (who knows, perhaps the latest liveCDs have it also). The only advantage of using this I can see is the ability to to make use of the BIOS RAID helpers to create and manage your arrays, and deal with the inherent boot time issues. I'm not sure about the stability or reliability of this though, as I havn't used it, and the readme doesn't really give me courage :-). Anyone using this successfuly ? It seems interesting. I just put together a little home server which uses both Linux RAID (md) and LVM2 on an old Abit KG7-RAID motherboard. Even though it has a built in Highpoint HPT37X "RAID chip" (a ghetto-RAID BIOS helper), I decided to go with good old "md". I've tested it by pulling power on drives, and it even boots up when the 'primary' drive doesn't exist (boot blocks on both mirrored drives of course). Seems to work very well. I have /boot mirrored (md0), and root and swap on LVM2 partitions which live on another mirrored partition (md1). For any wanting to do similar, you just need to set up GRUB on both drives, and make sure your have initramfs support for starting up md and LVM2. Generkernel will produce a kernel with this if you compile the md drivers into the kernel, and include --lvm2 in the genkernel flags. Make sure you include "dovlm2" and lvmraid=/dev/mdX lines for each of of your RAID devices on the boot line, which tells the linuxrc scripts to start up your RAID devices in the initramfs so it can mount your LVM2 root partition. - Jim -- +-----------+ | Jim Burwell - Sr. Systems/Network/Security Engineer, JSBC | +---+ | "I never let my schooling get in the way of my education." - Mark Twain | | "UNIX was never designed to keep people from doing stupid things, because | | that policy would also keep them from doing clever things." - Doug Gwyn | | "Cool is only three letters away from Fool" - Mike Muir, Suicyco | | "..Government in its best state is but a necessary evil; in its worst | | state an intolerable one.." - Thomas Paine, "Common Sense" (1776)| +---+ | Email: [EMAIL PROTECTED] ICQ UIN: 1695089 | +---+ | Reply problems ? Turn off the "sign" function in email prog. Blame MS. | +---+
Re: [gentoo-user] Knoppix Install Method
though. -- +---+ | Jim Burwell - Sr. Systems/Network/Security Engineer, JSBC | +---+ | I never let my schooling get in the way of my education. - Mark Twain | | UNIX was never designed to keep people from doing stupid things, because | | that policy would also keep them from doing clever things. - Doug Gwyn | | Cool is only three letters away from Fool - Mike Muir, Suicyco | | ..Government in its best state is but a necessary evil; in its worst | | state an intolerable one.. - Thomas Paine, Common Sense (1776)| +---+ | Email: [EMAIL PROTECTED] ICQ UIN: 1695089 | +---+ | Reply problems ? Turn off the sign function in email prog. Blame MS. | +---+ -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] raid1 + lvm2
Jarry wrote: Hi, I'm going to re-install gentoo on a small hobby-server and because I need both redundancy and flexibility, I thought in addition to raid1 (2x 160GB ata-disk) this time I would also use lvm2: /dev/md0 /boot (~50MB) /dev/md1 / (2GB) /dev/md2 swap (2GB) /dev/md3 lvm2(rest for /var /tmp /usr /opt /chroot /home) Is this generaly advisable solution (lvm2 over raid1), or is there some risk in using raid1 together with lvm2 ? One more question concerning partition type: If I want to use raid1, I have to set all those 2x4 (hda+hdc) primary partitions as type fd (raid autodetect). Is it not problem later for lvm2 when preparing and creating volume-group? Because lvm-guide says something about setting partition type to 8e (linux lvm), which I can not do, if I want to use raid1... Jarry I recently set up a server like this. I have /boot, swap, and root mirrored using linux RAID (md), and swap and root partition is an LVM2 partition. There's no problem with setting the partition types of 0xfd. LVM2 doesn't have a problem with it. One thing you need is a initrd or initramfs setup to get all this stuff up and running during boot. I found the easiest way to do this was to use genkernel. Here's are some quick notes on how I got this working using the gentoo-sources-2.6.14-r2 kernel: Create your partitions using fdisk (/boot, and LVM2 partition for swap and root) on both disks. Set up RAID1 mirroring for these partitons (/boot md0, LVM2 swap/root md2) Add the md1 device as a physical volume, and create a volume group and logical volumes for swap and root inside it. Do high level formats on your new /boot partition and root volume, and mkswap on your swap volume.. Use genkernel to configure your kernel making sure to include this option: --lvm2 Make sure the linux RAID (md) stuff is compiled into the kernel. The LVM2 (dm) stuff can be modules. Make sure you include these kernel flags in your grub.conf or lilo.conf file: root=/dev/ram0 init=/linuxrc real_root=/dev/vg00/root dolvm2 lvmraid=/dev/md0 lvmraid=/dev/md1 (Those last two flags don't seem to be documented anywhere, but are required for the genkernel produced linuxrc scripts in the initramfs to start up the md devices before it scans for the LVM2 stuff) All this is probably easier done, if possible, during the install from a livecd. I did it backwards myself. I got a basic gentoo system running on one drive with normal partitions, then transformed it into the RAID/LVM system. I did this by creating the Linux RAID partitons on the second drive with 'missing components' using mdadm (I use mdadm instead of the traditional raidtools). This allows you to create your RAID1 volumes with only one drive, the volumes coming up in degraded mode. You can then do your LVM2 volume creation stuff and create your filesystems, etc, get them mounted, and copy your basic system from the non RAID drive to your new volumes. Then you can update your fstab, set up your bootloader on the RAID drive, and test boot, etc, making sure the system comes up using the RAID/LVM2 setup, while having the original drive for backup in case things don't work. Once you're confident things will come up on the RAID/LVM2 setup, you can repartition the original drive, and "hotadd" the partitions to your RAID volumes. Linux RAID will then sync the formerly missing components, and voila, you now have a synced/clean RAID1 system. Hopefully that all made sense. I banged my head against it for a while getting this set up to boot up (the final key was the lvmraid= flags which cause the linxrc to start up the md devices. I had to read through the linuxrc scripts to figure this one out). Good luck. -- +---+ | Jim Burwell - Sr. Systems/Network/Security Engineer, JSBC | +---+ | "I never let my schooling get in the way of my education." - Mark Twain | | "UNIX was never designed to keep people from doing stupid things, because | | that policy would also keep them from doing clever things." - Doug Gwyn | | "Cool is only three letters away from Fool" - Mike Muir, Suicyco | | "..Government in its best state is but a necessary evil; in its worst | | state an intolerable one.." - Thomas Paine, "Common Sense" (1776)| +---+ | Email: jimb at jay ess bee cee dot com ICQ UIN: 1695089 | +---+ | Reply problems ? Turn off the "sign" function in email prog. Blame MS. | +---+