IRC Logs for #crux-devel Wednesday, 2013-01-23

*** MadSkipjack has joined #crux-devel00:16
*** MadSkipjack has quit IRC00:16
*** MadSkipjack has joined #crux-devel00:17
*** MadSkipjack has quit IRC00:18
*** MadSkipjack has joined #crux-devel00:18
*** MadSkipjack has quit IRC00:19
*** MadSkipjack has joined #crux-devel00:19
*** MadSkipjack has quit IRC00:19
*** MadSkipjack has joined #crux-devel00:19
*** MadSkipjack has quit IRC00:20
*** MadSkipjack has joined #crux-devel00:21
*** MadSkipjack has quit IRC00:26
*** MadSkipjack has joined #crux-devel00:27
*** MadSkipjack has quit IRC00:27
*** MadSkipjack has joined #crux-devel00:27
Romsterhi01:45
juehi Romster01:46
Romsterhi jue01:46
Romstergot my wind screen wipers fixed, been lucky it hasn't rained much01:47
jueRomster: the gutenprint issue should be fixed now01:47
Romsterk i'l test after i do a kernel update.01:47
Romsteri'm getting sick of some locoking problem building in a chroot on top of a logical volume seems to end up in D state...01:48
Romsterjue, you able to add compat-32 to cruxbot? since i think tek is really busy lately.01:50
jueRomster: will try01:59
jueRomster: compat-32 should work now02:25
teK_hey I'm not 'busy' :]02:28
juehi teK_02:32
Romsterk02:34
Romsterhi tek02:34
teK_hey guys02:34
Romsternow to figure how to setup my /boot/extlinux.conf02:34
jueRomster: btw, the /dev/disk/by-id/... pattern works nice with lilo03:33
Romsteroh...03:41
Romsterwhy didn't i try that before...03:41
Romsterthere is no clear consise info on syslinux03:41
Romsterusr/sbin/lilo-uuid-diskid i haven't noticed that file before.03:42
jueyou can use lilo-uuid-diskid to convert the boot= device and use the same device-name with appended -part1, -part2 for the root= device03:44
Romsterhmm running lilo-uuid-diskid made03:47
Romsterboot = /dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN root = "UUID=597f780b-a917-400f-86a7-22a1a0efc9a0"03:47
RomsterWarning: /dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN is not on the first disk03:47
RomsterAdded CRUX-3.7.3-pf  *03:47
jueRomster: the UUID thing will probably not work03:48
Romsterpartiton 1 is root... but you don't set the partition in boot=03:48
Romsteryeah i was wondering that i think i need a initramfs for that to work03:48
Romsteri had sdc1 there before...03:48
juefor root= yu should use something like this:03:49
Romsterso if i just hack it to root = /dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN-part103:49
jueroot=/dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN-part103:49
jueyep, that works03:49
Romsteri never did try that, thank you for the tip...03:50
juenp03:50
Romstersometimes the simplest things are best... honestly the other boot loaders are a pain i far better like lilo.03:50
Romsteris not on the first disk, ah that's just the BIOS disk order i should probably reorder that i guess.03:51
Romsterrebooting03:52
*** Romster has quit IRC03:52
*** horrorStruck has quit IRC03:59
*** Romster has joined #crux-devel05:20
*** Romster has quit IRC05:20
*** Romster has joined #crux-devel05:20
Romstersysuped my system... every time i reoroder the disks on the controller i get a kernel panic http://imagebin.org/24392305:22
Romsterhttp://pastebin.com/ZNH0Ekut lilo.conf05:23
Romsteri expect i should be able to shove the SSD on any sata cable and as long as it's the first one listed in the BIOS boot order it should boot and work.05:24
jaegerThat's usually the case on modern chipsets05:26
jaegerI can do that with my AHCI controllers without trouble, for example05:26
jaegeras long as UUIDs are used instead of raw devices05:26
jaegerno idea on the 'by-id' method, never tried that05:27
jaegerI also don't use lilo05:27
Romsteri was on grub2 didn't do what i want (same deal with re-ordering) and it's configuration  is O_O crazy messy.05:28
Romstersyslinux is stupid.05:28
Romsteri have to mount /boot to even get it to install.05:28
Romsterwhat are you using jaeger ?05:29
jaegerI use grub2 on most installs05:29
jaegerI don't use grub-mkconfig, though, my configs are fairly simple05:29
Romstersigh i already nuked that off my system...05:29
Romsterwhy can't disk/by-id not work05:30
Romsterseems to hard code it still to sdc105:30
jaegerIt seems like it IS not working :D05:30
Romstermy attention span and pissed off state makes me wanna rant for hours on the damn thing but i'll just go do other stuff and try again later... guess i'll just make a damn initramfs.05:31
Romsterheck is there a lilo channel05:32
Romsterfigures #lilo is unavailable...05:33
jueRomster: I don't see any error in your lilo.conf, a similar config works for me05:34
Romsteri've had this issue for ages just now it's getting on my nerves again.05:35
Romsteri also can't see why it would not stop working it's the same disk controller...05:36
Romsterjust different ata channel.05:36
Romsterand i do go into BIOS to make the SSD as the first item to boot from.05:36
Romsterwhy would changing ata channels make any difference.05:36
jaegerI'm confused, I thought you said it did stop working... now you say you can't see why it would not stop working05:37
jaegeror was the negative unintentional?05:38
juedo you get a different dev/disk/by-id/ata-....  if you change the ata channel of the disk?05:38
Romsteri have 2 controlelr one with 6 ports one with 2 ports. i got two raid1s setup in mdadm. one leg of each of the raids is on that other controller...05:38
jaegeralso, does it work with UUIDs instead?05:38
RomsterUUID's wont work on lilo without a initramfs05:38
Romsteri couldn't get grub2 to work with UUID's and i failed to figure it's syntax out than using the grub-mkconfig05:39
jueRomster: ?05:39
Romsterand my /etc/fstab already has UUID's in it so that cant be the problem.05:40
jueRomster: again, do you get a different /dev/disk/by-id/ata-....  if you change the ata channel of the disk?05:41
Romsternever mind i'm in a annoyed state at why this wont work. every attempt on and off months fails.05:41
Romsterno idea why would that change?05:41
juedid it change ?05:41
Romsteri can't even get in due to kernel panic to see... well can you tell by the panic image i posted?05:42
Romsterby-id is not meant to change.05:42
jaegerI don't think you need initramfs for lilo to use UUIDs05:43
jaegerits manpage doesn't explicitly say that, at least05:43
Romsterhttps://wiki.archlinux.org/index.php/Persistent_block_device_naming u'll read that another time when i'm not in a crappy mood.05:44
Romsteri'll*05:44
juejaeger: using UUID for root= dosn't work05:45
RomsterTo use persistent names in your boot manager, the following prerequisites must be met:05:46
Romster    You are using a mkinitcpio initial RAM disk image05:46
Romster[...]05:46
Romsterneather does05:46
Romsterboot=/dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN05:46
Romsterroot=/dev/disk/by-id/ata-INTEL_SSDSC2CT120A3_CVMP229400TL120BGN-part105:46
Romsterit still hard codes it as sdc1 in my boot loader.05:46
Romsterwhich it is sdc1 currently i'd hve to boot a live cd chroot and run lilo to make it hard coded to the new ata location...05:47
jaegergrub2 makes this pretty easy for me05:47
Romsterthe whole point of this is to make it independent and bot loader to find the disk on it's own.05:48
RomsterIf you are using LILO, then do not try this with the root=... configuration option; it will not work. Use append="root=..." or addappend="root=..." instead. Read the LILO man page for more information on append and addappend.05:48
Romsterhmm i didn't notice that...05:48
Romsterhonestly why can't any boot loader scan all disks to see what disk it's on... then read the lilo.conf to find what root is on.05:50
RomsterUsing a root specification of the form root="UUID=xxx..." or the form root="LABEL=xxx..." causes LILO to pass that literal string to the kernel at boot time in the kernel boot parameters.  For example,05:53
RomsterHowever, in order for the kernel to be able to successfully translate this into a specific partition, such as /dev/sda2, your kernel must have an initial RAM file system, udev and its dependencies must be in the initial RAM file system, and udev must have been previously started with enough of a profile to enable it to find all the disk devices and their partitions, and to have read all the UUIDs and LABELs from all the partitions and to create05:54
Romster symbolic links for them in the directories under /dev/disk.  Otherwise, the kernel will not be able to find the root file system.05:54
Romsterhttp://users.wowway.com/~zlinuxman/lilo.htm05:54
Romsteri've read similar from other sites.05:54
Romsterhence why i tired grub2 it's meant to be able to do UUID without a initramfs but it's config is ugly i failed to understand it... jaeger if you got a sample that works mind pastebinning it?05:55
jaegerhttp://jaeger.morpheus.net/linux/crux/files/grub2-efi.cfg.txt <-- will start from this one and trim it a bit, give me a sec05:55
Romstersomeone else suggested syslinux so i tried that being simpler than grub205:55
Romstereither i'm too dumb for linux or i just have a very unusual use case.05:56
jaegerI can't offer anything for syslinux, never used it for things other than USB or CDs05:57
Romsterthe issue is this damn udev and kernel not being one.05:57
Romsterso it makes the kernel not able to resolve UUIDs at all without udev help in a ram disk.05:57
jaegerhttp://pastebin.com/LXNVYhhU05:57
jaegerI believe that's about as minimal as it can get and still work05:58
jaegeryou should also be able to change the root=05:58
Romsterit's not like i'm asking how to assemble a raid array and luksOpen the raid then activate lvm on top of the open crypto container.05:58
jaegerI haven't needed to do that in a long time05:59
Romsteri stick that in menu.1st correct?05:59
jaegerhttp://pastebin.com/kb3k3XqE06:00
jaegeruse /boot/grub/grub.cfg, menu.lst is an older grub 0.9x thing06:00
Romsterand i'll have to install grub2 again and install that into /dev/sdc06:00
jueRomster: I think the problem is that the kernel doesn't understand what a UUID is, I guess you have to use a initrd with any bootloader if you want UUID's06:00
Romstersearch --fs-uuid -set root yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyy <- that does not the UUID= bit in that line?06:01
*** horrorStruck has joined #crux-devel06:01
Romsteri read grub2 doesn't need a initramfs.06:01
jaegerI don't understand the question06:02
Romsteron line 5 why does it not have root UUID=yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyy06:03
Romsterdoes it assume it's a UUID already?06:03
jaegerthat's what --fs-uuid does06:03
jaegergrub2 has multiple search modules, that's the UUID one06:03
Romsterah06:04
Romsterok understood now, i think my problem with grub2 is i barely understand it.06:04
jaegerAnd I could be wrong about it working with UUIDs without an initrd, maybe06:04
jaegerIt's been years since I tried that06:05
jaegerbut I *think* it works06:05
jaegerwell, I guess in this case it would be a question of whether or not the kernel supports it, not the boot loader06:05
juewell, I doubt that it works, it might work with GPT06:05
jaegerbecause grub obviously supports it with the search.fs_uuid module, no idea on the kernel06:05
jaegerit's the kernel that has to deal with the root=UUID=blah on the command line06:05
jaegerthe UUID should be the filesystem's UUID, though, not specific to msdos vs. GPT labels, right?06:06
jueyeah, but the kernel doesn't understand it, IMO06:06
jaegersome docs in the gentoo wiki also say I'm wrong so I probably just don't remember it correctly, sorry06:07
Romsterafaik grub2 was made to remove the initramfs requirement do use it's grub2 modules as helpers.06:08
jaegerwell, I can test it here when I get to a point where I can reboot, working on something at the moment06:08
jueRomster: as long as the kernel cannot use the UUID's that will not work06:09
Romsterno rush i'm gonna goto bed shortly06:09
Romsteri'm positive ir ead about the grub2 and UUID not requiring a initramfs either in the ML or on the grub2 site itself... but due to me doing far more stuff at work and in my own time i mix everything up...06:10
Romsteri honestly think i'm trying todo too much and i've found my limitations.06:11
Romsterfresh mind another day, i'll go do more reading.06:11
Romstersorry to take your time guys. enjoy your days.06:14
juenot sure, but IMO you have to use GPT partitions to achieve what you want06:14
juemeaning without an initrd06:15
jues/initrd/initramfs/06:17
Romsteroh hmm you could be right there.06:18
jaegerok, tested it here, it didn't work06:19
Romsteri didn't ting it worth while making a 120GB SSD as GPT06:19
jaegerI went and looked at kernel-parameters.txt and init/do_mounts.c in the kernel source, nothing there06:19
Romsters/ting/think06:19
Romsteri'm trying todo the impossible..06:19
Romsterhow does windows do it so fluently.06:20
jaegerIt might be best just to make an initramfs and not worry about it06:20
Romstercontroller(0)\disk(0)\partition(0)06:20
Romsteryeah all for a appears to be simple todo...06:21
Romsteroh well learning up on initramfs.. another day.06:21
jaegersorry for the noise, I thought it could be done for some reason06:22
Romsterit's fine i thought i read grub2 supported UUID without a initramfs... i can remember reading that somewhere.. can't find where now.06:23
jaegeron a side note it looks like I didn't finish the unifont install in the grub2 port, oops06:23
Romsterthe whole boot loader and kernel stuff is a royal pain.06:24
Romsterhehe06:24
jaegerIt is, yeah, and I get extra confused sometimes due to making EFI bullshit work06:30
jaegeractually planning to convert my laptop back to BIOS/msdos from EFI/GPT today06:32
jaegerIt was an interesting project but not really all that useful and now I can test EFI stuff in VMs06:32
Romsteri was reading what you did seems too painful06:52
jaegerIt was pretty rough initially but I learned a lot, it's easier now06:52
Romsterlearning experience yeah always good to have.06:57
Romsterjust like i've learned more about boot loaders.06:58
frinnstmy drives are gpt partitioned, but no efi stuff (i think :D)07:33
frinnstused the gptmbr.bin in syslinux07:33
jaegeryeah, that's not unusual now07:34
jaegerEFI requires GPT but GPT does not require EFI07:34
*** joe9 has quit IRC09:06
*** joe9 has joined #crux-devel09:07
*** joe9 has quit IRC11:14
frinnstvmtouch looks like a nice tool12:59
*** sepen has joined #crux-devel13:00
*** sepen has quit IRC13:00
jaegerneat, never heard of that one13:02
frinnstme neither13:02
frinnsti had 17% of linux-3.8-rc4 in mem :D13:03
*** sepen has joined #crux-devel13:07
sepenhmm I have perms to change BRANCHES in hook/post-update, should that be enough13:08
sepenhmm no13:11
*** sepen has quit IRC13:12
*** sepen has joined #crux-devel13:12
*** sepen has quit IRC13:12
*** mike_k has quit IRC13:22
*** joe9 has joined #crux-devel17:07
*** acrux has quit IRC19:35
*** __mavrick61 has quit IRC19:39
*** __mavrick61 has joined #crux-devel19:40
rmullHow come we don't just use git to distribute ports, instead of httpup and rsync? Free history for ease of downgrade19:41
prologicWe do19:58
prologicand mercurial too19:58
prologicor rather we can19:58
prologicbut it doesn't buy us much ihmo19:58
prologicexcept better collaboration19:59

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!