You are not logged in.
Pages: 1
I did a pacman -Syu today and part of the update was the linux package. Unfortunately, it froze on
-> Parsing hook: [autodetect]
After several hours, I killed the upgrade process and did
pacman -S linux
It froze again in exactly the same place. I'm now somewhat nervous of rebooting my machine, since I'm not sure that the kernel is installed properly... How can I get passed this and have a properly booting and upgraded machine?
Last edited by phunni (2011-09-01 07:19:48)
Offline
Perhaps you haven't changed the grub menulist to reflect the new kernel ID as described in archlinux home page news.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
If you mean the changes from when the package became linux (http://www.archlinux.org/news/changes-t … filenames/), then I have. Did it a while ago.
Offline
Have you tried the fallback?
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
I haven't tried rebooting at all - I'm too scared to be honest! Obviously, if there's no other option, I'll have to. I'm rather hoping, though, that there's a way of successfully completing the upgrade.
I noticed that there were a number of mkinitcpio processes - presumably from all my attempts. I tried killing them all incase that would free it up, but I got exactly the same result.
Offline
I've also tried clearing the cache and redownloading the package - but the result is the same...
Offline
Currently downloading an older version of the kernel to see if a downgrade will work...
Offline
Nope - same result... :-(
Offline
You can also boot from live media and rebuild the kernel. See the "install existing" article on the wiki. How about trying a totally different kernel package like kernel26-lts?
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Maybe the hardware time is wrong?
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Several thoughts:
Edit: Is your mkinitcpio up to date? There used to be a bug that caused it to hang like this. If you've got an upgrade to mkinitcpio coming in with your kernel upgrade, upgrade mkinitcpio first, and then do the kernel with the new mkinitcpio.
The autodetect hook runs the code in the build function of /lib/initcpio/install/autodetect, which in turn runs the auto_modules function from /lib/initcpio/functions. Look at these and try to figure out what in there could be hanging.
Does "mkinitcpio -M" also hang? It runs the same autodetect function (but with different output).
Check what child processes (if any) the hung mkinitcpios have before you kill them. (pstree may help with this)
Don't forget that /sbin/mkinitcpio itself is actually a shell script. Change the first line of it (temporarily) from just "#!/bin/bash" to "#!/bin/bash -x" and you will get gobs of debugging info when mkinitcpio runs.
You don't have to actually reinstall the kernel package again to retry mkinitcpio - you can just do "mkinitcpio -p linux", which is what pacman runs for you on kernel upgrades.
Last edited by ataraxia (2011-08-31 22:53:00)
Offline
Thanks for the help! It appears to be getting stuck running the following:
/sbin/mdadm -Esv /dev/sda /dev/sda1 /dev/sdb /dev/sdb1 /dev/sdc /dev/sdc1 /dev/sdc2 /dev/sdd /dev/sdg /dev/sdg1 /dev/sdg2 '/dev/ida/*' '/dev/cciss/*' '/dev/ataraid/*' /dev/mapper/control /dev/mapper/vg-syshome /dev/mapper/vg-sysroot /dev/mapper/vg-sysswap
According to ps, there are several mdadm processes running. I have tried killing them with 'killall -9 mdadm' but they won't die...
Offline
BTW -I'm not using RAID at all, but I am using LVM2
Offline
Are you running RAID on your system?
If not it looks like you are running a module that is unneeded, if that is the case wouldn't rc.d stop mdadm kill it? (as root of course)
Offline
I'm not running RAID. I tried the process outlined here: https://bbs.archlinux.org/viewtopic.php … 84#p542584 and it has worked. Still a little bit nervous about a repeat since I still don't know the cause, but at least it's working for now...
Offline
Glad you got it working.
Seems odd though that autodetect found mdadm initially. I'd go though all the configuration files and see if I could find where it's getting loaded & disable it if it were me, but maybe it's run by default? Which doesn't make sense, I run Raid and I had to edit the mkinitcpio.conf myself to include mdadm.
If you do truly solve the problem I'd like to know how you did it.
Offline
Pages: 1