Andrei Borzenkov wrote:
On Thu, Apr 20, 2017 at 8:04 AM, L A Walsh <suse@tlinx.org> wrote:
Andrei Borzenkov wrote:
19.04.2017 04:38, L A Walsh пишет: Why do you expect them to autoload? Most needed modules autoload as I use them on my sysV system that boots from the hard disk.
I almost hit DEL ... yet again you start blaming systemd for every dead kitten.
Woe woe woe... 1) When have I ever blamed anything on sysd? I've never run it, so how can I ever have blamed anything on it? 2) I didn't mention sysd. I only said that most modules autoload as needed on my sysv system that boots from the hard disk.
To make myself clear - your question sounded like you *know* that there is mechanism to autoload DM modules and this mechanism stopped working.
But its true that most of my modules are loaded automatically. I just rebooted my main system. By the time I was able to log in and run "lsmod", 43 modules had loaded -- including a module for dm-thin.
I myself am not aware of this from kernel side, so I do not see why you expect *autoload*.
It very likely isn't the kernel but something in the way the system is started and/or the fact that all of the modules are available to load as soon as the root file system is mounted. I.e. I don't have to worry about whether or not something was on an initrd or not.
That something loaded it in the past for you (or that you probably used your own monolithic kernel that did not need modules) does not mean it was *autoloaded*.
---- You must be thinking of someone else. I don't have a monolithic kernel, to the contrary, the suse kernel is about 50% larger than mine (9M vs. 6M). As far as modules go, I have 539 dynamic modules that are loadable as the system comes up and later, as needed. So far, you've confused me with someone who's run sysd, and someone who has a monolithic kernel, when neither are even remotely true. ---- The real problem, so far has been my using thinpools, which *are* supported in YaST, but are broken. FWIW, if I mod-probe the dm-thinpool module, the 'vgchange -a y' installs all of the devices needed for thinpool device.
Yeah, I must have been confused or did not look there carefully for a long time. Sorry. Indeed, YaST offers thin pool - which makes this issue topic of bug report. I am not sure what component though - it is not exactly kernel bug as far as I can tell (it could be considered missing feature) so options are to either load modules from DM library or statically configure them for loading by YaST when thin pool is created. Given that currently even simple "lvcreate -T" fails, I'd vote for the former.
lvcreate -T fails? Worked for me (w/the modules loaded). I don't like sysd because it leaves me with a non-bootable system. You can't claim that's unfair when its the truth. As for your other remarks, you have to be thinking of someone else, as neither of your comments _could_ apply to me. I still am not sure how to keep my VM "up", since I'm not sure how to pre-load the needed modules for LVM, (putting modules in /etc/modules-load.d/ as some suggest doesn't work). I did figure out howto mount my main system's root disk, so as a last resort, I can always download a working boot from it, but I really wanted to setup a server VM without my own custom solutions.... *cheers*, and don't start hitting 'DELETE' until you are sure you've got the right person.... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org