tuxd3v | hello, does any one that is already in beowulf is noticing swap not been mounted at boot time? | 01:48 |
---|---|---|
tuxd3v | even with 'sw' directive? | 01:48 |
rrq | the fstab line is fine | 01:56 |
rrq | ? | 01:56 |
rrq | ie right partition, with right formatting? | 01:57 |
tuxd3v | yup | 02:00 |
tuxd3v | migrated from ascii to beowulf.. | 02:00 |
tuxd3v | and no more swap | 02:00 |
tuxd3v | need to mount it by hand.. | 02:00 |
aaro | maybe the uuid changed? | 02:03 |
tuxd3v | no.. | 02:04 |
tuxd3v | its zram :) | 02:04 |
rrq | md5sum /etc/init.d/mountall.sh = 0850a5381c8d8c356df53a2b64c69388 | 02:04 |
rrq | which arch btw? | 02:04 |
tuxd3v | aarch53 | 02:07 |
tuxd3v | armhf | 02:07 |
tuxd3v | aarch32 | 02:07 |
tuxd3v | its the same checksum :( | 02:09 |
tuxd3v | thanks | 02:09 |
rrq | ok. and sysvinit? that should run /etc/init.d/mountall.sh which should do swapon .. unless boot option noswap | 02:09 |
rrq | (my md5sum is beowulf/i386) | 02:10 |
tuxd3v | if I do by hand 'service mountall.sh start' | 02:11 |
tuxd3v | is starts | 02:11 |
tuxd3v | the same as with swapon -a | 02:11 |
tuxd3v | but not automatically | 02:11 |
tuxd3v | its strange | 02:12 |
rrq | and it's enabled in /etc/rcS.d/ ? | 02:12 |
tuxd3v | maybe that is the problem | 02:12 |
tuxd3v | I will see ;) | 02:12 |
tuxd3v | lrwxrwxrwx 1 root root 21 Nov 16 03:28 /etc/rcS.d/S08mountall.sh -> ../init.d/mountall.sh | 02:16 |
tuxd3v | only apear after: | 02:16 |
tuxd3v | https://paste2.org/6YaeEpB3 | 02:17 |
tuxd3v | my fstab: https://paste2.org/cAFHxUI9 | 02:19 |
tuxd3v | its dificult to explain why.. | 02:19 |
tuxd3v | If forced by hand ok, it comes up | 02:20 |
tuxd3v | but not on boot | 02:20 |
tuxd3v | even with 'sw' option.. | 02:20 |
tuxd3v | One thing that I noticed is that on transition from ascii to beowulf | 02:20 |
tuxd3v | some files become a bit unformated | 02:20 |
tuxd3v | I needed to align code by hand | 02:21 |
tuxd3v | udev rules from zram creation were an example | 02:21 |
tuxd3v | fstab is another. | 02:21 |
tuxd3v | the information was there | 02:21 |
tuxd3v | but the '\n' at end of last line simply disapeared | 02:22 |
aaro | no errors in dmesg? | 02:22 |
tuxd3v | And this was how I found out that my swap is not auto mountable anymore :) | 02:22 |
tuxd3v | at least the finfo related with swap seems ok | 02:24 |
tuxd3v | https://paste2.org/a6BApALz | 02:24 |
tuxd3v | a 'mount -o remount -a' | 02:25 |
tuxd3v | only a service mountall.sh start | 02:25 |
tuxd3v | or swapon -a | 02:25 |
tuxd3v | solves it.. | 02:25 |
tuxd3v | I know that sysctl syscal will be dropped I believe I don't know if kernel 5.5 or even was in 5.4.. | 02:26 |
tuxd3v | I don't find nothing unusual.. | 02:29 |
tuxd3v | it could be a timeout on zram creation and formating at boot time? | 02:30 |
tuxd3v | unrelated, but maybe related :S | 02:31 |
tuxd3v | http://manpages.ubuntu.com/manpages/xenial/man5/systemd.swap.5.html | 02:31 |
tuxd3v | TimeoutSec= | 02:31 |
stovepipe | what changed to make ls on a dir with links to automounted volumes force them all to mount instead of just listing the dir with links | 02:32 |
stovepipe | maybe not devuan specific, but is that standard linux behavior now? | 02:33 |
tuxd3v | stovepipe, I never saw ls mount volumes | 02:34 |
stovepipe | it is unpleasant | 02:35 |
stovepipe | maybe an autofs config thing | 02:35 |
fsmithred | it shouldn't mount until you look inside that particular directory | 02:35 |
stovepipe | i'm still looking | 02:35 |
stovepipe | yeah its odd | 02:35 |
fsmithred | but I've never tested it in ascii | 02:35 |
stovepipe | i have a map for /hosts that is where all my nfs mounts are, and link there in various places | 02:36 |
tuxd3v | stovepipe, do you mean via loosetup? | 02:36 |
stovepipe | just ls forces a mount | 02:36 |
tuxd3v | a loop device? | 02:36 |
fsmithred | no, he means autofs | 02:36 |
stovepipe | of any dir with a link to a mount point | 02:36 |
stovepipe | i had 2.0 on another machine and i dont remember it doing that, but that install is gone onw | 02:37 |
stovepipe | this is 2.1 from the latest spin | 02:37 |
stovepipe | the other one is fast and i might not have noticed | 02:38 |
fsmithred | take me a few minutes, but I can try it on the install I did yesterday | 02:38 |
stovepipe | i was playing with antix on this same machine, and i might have noticed it there also, but i wasnt making it permanent so i didnt think about it | 02:40 |
fsmithred | yup. That's fucked up. | 02:54 |
fsmithred | good thing I don't have a dozen links to automounts sitting in my home dir | 02:55 |
stovepipe | lol | 02:55 |
fsmithred | I'll check beowulf in a couple minutes | 02:55 |
tuxd3v | rrq, crazy.. still can't found a solution.. | 02:57 |
tuxd3v | I even tried to clone mountall.sh script atribute a diferent name create a new service to it.. but no swapon by default.. | 02:58 |
tuxd3v | but If I create a ascii image right now, I know it will work, then on upgrading, zram puff.. | 02:59 |
tuxd3v | its created 'lz4' compression algo attributed | 02:59 |
tuxd3v | zram kernel module is loaded | 02:59 |
tuxd3v | but no swapon | 02:59 |
tuxd3v | even using 'sw' as parameter in fstab.. | 03:00 |
rrq | do you have bootlogd installed? try uninstalling that ... I have a vague memory of something about that | 03:01 |
fsmithred | stovepipe, it does it in beowulf, too | 03:01 |
rrq | fstab should be like "/dev/sda3 swap swap defaults 0 0" though the first "swap" is irrelevant | 03:03 |
rrq | the partition may alternatively be named by label or uuid as well of course | 03:04 |
fsmithred | I use 'none' | 03:04 |
fsmithred | for swap mountpoint | 03:04 |
rrq | but fstab must be ok since it works by hand | 03:06 |
tuxd3v | yep | 03:06 |
tuxd3v | doing a swapon -a | 03:06 |
tuxd3v | works | 03:06 |
tuxd3v | or running /dev/init.d/mountall.sh | 03:07 |
tuxd3v | but by hand because on boot | 03:07 |
tuxd3v | no way.. | 03:07 |
tuxd3v | crazzy thing.. | 03:07 |
rrq | I remember a problem with the boot sequence, which made it exit prematurely, and not run mountall.sh ... but I dont remeber what it was :( | 03:07 |
rrq | if you have console, you should add some output to the beginning of each init file for tracking donw how far it gets | 03:08 |
rrq | I *believe* I resolved mine by not installing bootlogd | 03:09 |
rrq | which was a new thing with beowulf cf ascii | 03:09 |
rrq | (came automagically with debootstrap) | 03:11 |
fsmithred | I have bootlogd installed in beowulf and my swapfile gets mounted | 03:19 |
fsmithred | or used, I should say | 03:19 |
fsmithred | I don't have a beowulf with a swap partition right now | 03:19 |
rrq | mmm the other maybe relevant thing I can notice with my arm* builder debootstrap scripts is that they add mawk to the includes. | 03:25 |
tuxd3v | yeah mawk was present I substituted it, but maybe he left some incompatibilities.. | 03:37 |
tuxd3v | Now I am thinking in the unformated lines of udev rules and such.. | 03:38 |
tuxd3v | indeed I also found some incompatibilities related with strings.. | 03:39 |
tuxd3v | https://stackoverflow.com/questions/21289110/the-differences-between-gawk-and-mawk-column-width | 03:39 |
stovepipe | fsmithred: lol i just walked another person through dl'ing an iso and writing it with rufus, on a T510 identical to mine, and it has the same firmware problem from before | 03:47 |
fsmithred | yesterday's iso? | 03:47 |
stovepipe | with a new 2.1 image from leaseweb lol | 03:47 |
fsmithred | maybe a rufus problem | 03:48 |
stovepipe | err, no that one was from ksx4system | 03:48 |
fsmithred | I've heard of problems with using rufus, but I don't know any details | 03:48 |
stovepipe | i just installed 2.1 with a new image on a T510 last night with image from leaseweb | 03:48 |
stovepipe | had him write it again from that mirror | 03:48 |
stovepipe | same shit again lol | 03:48 |
stovepipe | i'm giong to test it now downloading and writing usb with rufus all on windows like he did | 03:49 |
fsmithred | in root of iso, /firmware should contain all symlinks | 03:49 |
fsmithred | maybe those aren't being copied | 03:49 |
stovepipe | rufus also has iso and dd mode | 03:57 |
stovepipe | this guy was never part of any of that other stuff either | 04:04 |
stovepipe | worked for me in dd mode | 04:04 |
stovepipe | having him try again | 04:04 |
fsmithred | yeah, I used dd and the new iso worked fine | 04:04 |
stovepipe | i dont see how an error with a tool like rufus could result in the same damn error lol | 04:05 |
fsmithred | like I said - maybe it doesn't copy symlinks | 04:05 |
stovepipe | lol it worked | 04:09 |
fsmithred | stovepipe, see if you can figure out a way to do it wrong | 04:22 |
stovepipe | use iso mode i guess | 04:22 |
stovepipe | i'll try it | 04:22 |
fsmithred | is iso mode for making a multi-boot live usb? | 04:23 |
stovepipe | rufus has two options at write time, iso mode and dd mode | 04:23 |
rrq | I think there is some story about Joliet vs Rock Ridge here... | 04:23 |
stovepipe | doing something else atm | 04:24 |
stovepipe | yeah and the links | 04:26 |
stovepipe | yeah it copies files from the image | 04:30 |
stovepipe | in "iso mode" | 04:31 |
fsmithred | puts them in their own directory? | 04:31 |
stovepipe | its copying individial files to the newly formatted usb | 04:32 |
stovepipe | instead of just blindly writing an image | 04:32 |
stovepipe | looks like it actually mounts it as a drive letter during the process | 04:33 |
fsmithred | how does it plan to boot? | 04:33 |
stovepipe | it writes all that also first | 04:33 |
stovepipe | takes its sweet time also | 04:33 |
stovepipe | ok here we go boot test | 04:34 |
stovepipe | not sure why they call it "iso mode" heh | 04:34 |
onefang | In Search Of a mode? | 04:35 |
stovepipe | and bzzzt | 04:35 |
stovepipe | rufus in iso mode cant handle links | 04:35 |
stovepipe | is the final verdict | 04:35 |
stovepipe | lol | 04:35 |
fsmithred | cool. now we know. | 04:35 |
stovepipe | what an annoying couple of days! | 04:36 |
stovepipe | when he read that error message to me i almost shit my pants | 04:36 |
stovepipe | NOT AGAIN!! | 04:36 |
fsmithred | lol | 04:36 |
fsmithred | two errors that lost the firmware | 04:37 |
fsmithred | maybe the universe is telling us not to use it | 04:37 |
stovepipe | and one tricky mirror | 04:37 |
rrq | in theory it's possible to copy into /firmware rather than link it | 04:53 |
stovepipe | fsmithred: browse_mode in /etc/autofs.conf | 07:57 |
fsmithred | stovepipe, thanks. That's weird that the default behavior changed from jessie to ascii, but the default setting did not. I changed 'no' to 'yes' and now it does what I expect. | 12:28 |
divansantana | are these repos correct? http://paste.debian.net/1122379 | 16:37 |
divansantana | seems I'm getting very little updates. | 16:37 |
divansantana | I'm getting Failed to load shared library 'libgdk-3.so.0' while trying | 16:40 |
divansantana | to launch virt-manager. Wondering is this because my setup is broken | 16:40 |
divansantana | or upstream? Guessing it's my setup. http://paste.debian.net/1122381 | 16:40 |
yeti | yay for orgmode! | 16:43 |
yeti | maybe " : Could not open display: localhost:11.0 " is the problem? | 16:43 |
yeti | I dont know virt-manager so just guessing | 16:43 |
Leander | possible, you might want to try "xhost +localhost" in a terminal to allow all local users to open windows in your current X session | 16:59 |
Leander | and maybe redo the sudo/su afterwards | 17:00 |
divansantana | yeti:lol. haha. And a bad paste. :) | 17:20 |
divansantana | Let me try that thanks. | 17:20 |
ukine | attn dev'rs i left a nice little winter wish in #debianfork. thanks again for maintaining/supporting this project :] | 18:18 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!