View previous topic :: View next topic |
Author |
Message |
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Wed Apr 23, 2014 12:59 am Post subject: Woes upgrading to Advanced Format Drive [SOLVED] |
|
|
I bought a WD250 Velociraptor to replace a WD320 RE drive. The new drive is Advanced Format with 4096 byte sectors. The old drive has 512 byte sectors and was cloned from it's predecessor with dd.
After booting sysrescuecd and creating the new partitions with 'parted -a optimal' I followed a tip from the web and copied the files with 'cp -afr /<old root> /<new root>' and the like for the boot partition. I then installed legacy grub per the handbook. When I rebooted, a grub prompt appeared. After hours of web searching and re-trials, I finally blew away the ext4 /boot partition and created an ext3 partition and copied files again. It got a little farther and I downgraded again to ext2. Now the partitions were the same type and the good work of 'mkfs.ext4 -b 4096' was gone for the boot partition. I went through the same steps of booting the CD, copying files and chrooting to install grub. Now when I boot I get to "waiting for uevents' (eudev, right?), a long wait, rapid lines with red stars in front, the screen flashes clear and 'This is (none)' appears on the screen. I went through the procedure AGAIN and edited /etc/rc.conf to be interactive. I then rebooted and started pressing the "i" key but the same thing happened. There is nothing for today in /var/log/messages that I can see from the chroot environment. I tried adding 'single_user=1' to the grub line but, you guessed it, the exact same thing happened.
Is legacy grub (0.97) having trouble with the /boot partition starting 1 MiB into the disk? Should I upgrade to grub2? I hate to because I don't understand the syntax and apparently after every kernel rebuild you have to run some sort of update program like we used to do with LILO. Is the Gentoo kernel having trouble with 4096 byte sectors? I tried running genkernel in the chroot and somewhere in busybox it got lost and stopped running without exiting.
Some one must have upgraded a drive from 512 to 4096. How did you do it, please? Should I re-install stage3 and a portage snapshot, copy my /root /etc and /home from the old drive, run "emerge -e --keep-going world", rebuild the kernel, reinstall grub and reboot? Last time I ran 'emerge -e world' it ran for 19 hours and that was on a faster machine.
Last edited by Tony0945 on Wed Apr 23, 2014 9:48 pm; edited 1 time in total |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21624
|
Posted: Wed Apr 23, 2014 1:40 am Post subject: |
|
|
I had no problems running Gentoo on a 4K sector drive, though that drive has since died. I aligned my partitions to 1MB. I believe Grub1 worked fine for that using ext3. If you got to the point of running udev, then your bootloader is no longer an issue. Something else is wrong with how you copied the system to its new drive. For each line with a red star, you need to fix the associated error. Start with the topmost failure. Sometimes a failure early on causes other failures later. Your hostname of (none) probably indicates a problem with /etc/conf.d/hostname. |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Wed Apr 23, 2014 2:19 am Post subject: |
|
|
I'll check /etc/conf.d/hostname. I can do that from sysrecuecd.
I can't read the red star items because they come too fast and then the screen blanks instead of scrolling. Is there anyway to fix that? |
|
Back to top |
|
|
DONAHUE Watchman
Joined: 09 Dec 2006 Posts: 7651 Location: Goose Creek SC
|
Posted: Wed Apr 23, 2014 5:14 am Post subject: |
|
|
suggest: boot a sysresccd, backup your files, fsck the gentoo partitions, mount the gentoo partitions,run the code's below looking for the quote's as a result Code: | grep -i "CONFIG_DEVTMPFS=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_DEVTMPFS_MOUNT=y" /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_DEVTMPFS_MOUNT=y |
Code: | grep -i "CONFIG_FSNOTIFY=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_DNOTIFY=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_INOTIFY_USER=y" /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_INOTIFY_USER=y |
Code: | grep -i "CONFIG_NET=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_PROC_FS=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_SIGNALFD=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_SYSFS=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_SYSFS_DEPRECATED is not set" /mnt/gentoo/usr/src/linux/.config |
Quote: | # CONFIG_SYSFS_DEPRECATED is not set |
Code: | grep -i 'CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"' /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" |
Code: | grep -i "CONFIG_BLK_DEV_BSG=y" /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_BLK_DEV_BSG=y |
Code: | grep -i "CONFIG_TMPFS=y" /mnt/gentoo/usr/src/linux/.config |
Code: | grep -i "CONFIG_TMPFS_POSIX_ACL=y" /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_TMPFS_POSIX_ACL=y |
Code: | grep -i "CONFIG_TMPFS_XATTR=y" /mnt/gentoo/usr/src/linux/.config |
Quote: | CONFIG_TMPFS_XATTR=y |
[/quote] recommend entering the chroot, editing menuconfig and recompiling and recopying the kernel if results are different. look for Quote: | udev | sysinit
udev-mount | sysinit | if not present add the entries.
dmesg should show the messages.
you may have lost a lot of symlinks in the copy process _________________ Defund the FCC. |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Wed Apr 23, 2014 3:25 pm Post subject: |
|
|
Thank you. The boot problem was due to a line in /etc/fstab that mounted a fourth partition as backup. Init went crazy when it couldn't mount the partition. I had already editted fstab to comment out mounting of the second (video) drive. I found this by recording the boot screen with a cell phone camcorder and playing it back, pausing just before the screen blanks. I mention this because it seems a usefull technique. The video, or still shots from it, can be uploaded to show the exact screen.
After fixing fstab, I was able to boot but couldn't login graphically because various hidden files were symlinked to themselves. command line login worked ok. I started deleting and copying from the command line. It seemed a monumental job. I'm now in the process of booting sysrescuecd (BTW, it comes up really fast from a usb stick) and using tar piped into tar to recopy the \home partition. I seem to have the syntax wrong and it's copying the root partition. It's been running for an hour and half and copying /usr right now. I'm going to leave the boot partition alone because it has no hidden files and is working. |
|
Back to top |
|
|
Chiitoo Administrator
Joined: 28 Feb 2010 Posts: 2573 Location: Here and Away Again
|
Posted: Wed Apr 23, 2014 8:16 pm Post subject: ><)))°€ |
|
|
Teegrins, Tony0945!
With regards to the screen being blankified, adding the --noclear option into /etc/inittab should change that.
Something like this:
Code: | # TERMINALS
c1:12345:respawn:/sbin/agetty 38400 tty1 linux --noclear |
Furthermore, within /etc/rc.conf lies the following:
Code: | # rc_logger launches a logging daemon to log the entire rc process to
# /var/log/rc.log
# NOTE: Linux systems require the devfs service to be started before
# logging can take place and as such cannot log the sysinit runlevel.
rc_logger="YES" |
I hope this helps a bit at least! _________________ Kindest of regardses. |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Wed Apr 23, 2014 9:38 pm Post subject: |
|
|
Yes it does, Chiitoo. In fact, I had the logger on, but I thought it logged to /var/log/messages. In fact, it logs to /var/log/rc.log. I saw that after I got it up and running, otherwise I could have examined the log from sysrescuecd. |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Wed Apr 23, 2014 9:46 pm Post subject: |
|
|
Found the right way to copy. Forget cp -a. Here's the actual commands that I gave, file permissions were preserved. Took about an hour and a half. He had a lot of large files. The other users only took about five minutes each.
Quote: |
root@sysresccd /root % mkdir /mnt/new
root@sysresccd /root % mkdir /mnt/old
root@sysresccd /root % mount /dev/sdb3 /mnt/old
root@sysresccd /root % mount /dev/sda3 /mnt/new
root@sysresccd /root % (cd /mnt/old/home/jacob && tar -cvpf - .) | (cd /mnt/new/home/jacob && tar -xvf -)
|
BTW, the ext2 partition that I made from gparted were optimal as later confirmed by parted. This consumed a lot of time, but hdparm shows its worth it. The old WD3200YS drive (made in May 2006! and still no errors) read at 65 MB/sec and the WD250HHTZ reads at 195 MB/sec! |
|
Back to top |
|
|
|