Call me kooky

Thanks to the server adventure I got to learn a lot about cloning linux drives. Normally I would just slap in a disk with ghost and be done with it, but I had a few issues this time. First, the source disk was in vmware. Second, the destination drive was in a USB enclosure and vmware only has USB 1 support.
I tried using dd at first and it took allllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll day. Part of that was the USB1 and part of it is that dd is very literaly a bit for bit copy. That takes forever of course.
Clonezilla doesn’t have USB support at all, so it’s out. I couldn’t find any other software that supported USB, not even ghost!
And that’s where g4u comes in. A nice little freebsd based iso that detects the USB drive and is much faster. I was getting 100k/sec with dd, that hopped up to 2mb/s with g4u. That was after I funked around with empty space though.
Direct link to the image reduction info.
And The main page.

Why can’t you just use simple flash instead of some fucked up software client that is a pain in the ass to install?

Dvorak on the spreadsheet. Always an enlightning read.

I decided to go for a walk and keep a track of it with my cell phone.  But for some reason the damned app couldn’t get a lock on any GPS sats.  No track at all.  Only thing I can think of is the overcast is mucking with it.  I’ll have to try again sometime.

First, I had to get a windows live id in order to do anything with the beta at all.  Stupid!  It would make more sense to make it easy to get the beta so you could, oh I don’t know, test the damn software?
Then I can’t download it in firefox at all.  I have to use IE and install yet another security hole/activeX componet.  It would just be easier to get the damned thing from bittorrent.

I ran fdisk real quick to check the md flag on the drives, which weren’t there.  Crap!  Then I forgot I had a nfs share mounted from a vmware machine.  Turning the vmware machine on allowd me to unmount that and give mounting the md discs a shot.

root@ubuntu:/# mount /dev/sdb mnt/oldserver/
mount: unknown filesystem type ‘linux_raid_member’

Whoops, forgot about that but we can force the fs type and check the data:

root@ubuntu:/# mount -t ext2 /dev/sdb /mnt/oldserver/
root@ubuntu:/# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda1              15G  6.8G  7.4G  49% /
varrun                569M  264K  569M   1% /var/run
varlock               569M  4.0K  569M   1% /var/lock
udev                  569M   52K  569M   1% /dev
devshm                569M     0  569M   0% /dev/shm
/dev/sda3              24G  6.9G   16G  31% /www
/dev/sdb              111G   69G   36G  66% /mnt/oldserver

Boink, one disk is still alive.

/dev/sdb              111G   69G   36G  66% /mnt/oldserver
/dev/sdc              111G   69G   36G  66% /mnt/dump

Both disks have the data correct data, yay.  Now we just need to rebuild the md array and such.

root@ubuntu:/mnt# mdadm
The program ‘mdadm’ is currently not installed.  You can install it by typing:
apt-get install mdadm
bash: mdadm: command not found
root@ubuntu:/mnt# apt-get install mdadm

Grk, first we have to install mdadm.  Why this isn’t included is beyond me.  it installed initramfs-tools also, whatever.  It might come in handy at some time.

Mdadm is a bit weird, and wants to create a brand new array.  Thankfully this works and doesn’t kill off the data.

root@ubuntu:/mnt# mdadm /dev/md0 –create -l 1 -n 2 /dev/sdb /dev/sdc
mdadm: /dev/sdb appears to contain an ext2fs file system
    size=117220736K  mtime=Wed Jan  7 17:54:26 2009
mdadm: /dev/sdb appears to be part of a raid array:
    level=raid1 devices=2 ctime=Wed Jan 16 19:46:19 2008
mdadm: /dev/sdc appears to contain an ext2fs file system
    size=117220736K  mtime=Wed Jan  7 17:57:41 2009
mdadm: /dev/sdc appears to be part of a raid array:
    level=raid1 devices=2 ctime=Wed Jan 16 19:46:19 2008
Continue creating array? y
mdadm: array /dev/md0 started.
root@ubuntu:/mnt# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sdc[1] sdb[0]
      117220736 blocks [2/2] [UU]
      [>....................]  resync =  0.1% (232832/117220736) finish=75.3min speed=25870K/sec

unused devices: <none>

Not sure about the error warnings, but after grinding away for a while it’s all up and working.

My whole fear was that creating a new /dev/md0 would wipe out the existing data, but it didn’t so yay me.  There’s nothing else to do after the resync, md is all set to go.

This thing is damn weird, some freaky stuff down there.
BBC article.

Finaly got the new os into the server…. sorta.  Turns out the PSU in the server puked and maybe took the board with it.  With one replacement cpu it reports a low 5v, with the other no 5v at all.  It refuses to boot at that point.  Fun.
So now the server is running off the old athlon system.  On the upshot it’s a lot faster with a lot more ram.  Been learning the ubuntu way of server setup and it hasn’t all gone smoothly.  Just got PXE booting to work correctly, whichis a bear since the PXE rom on the system I use is buggy to start with.
Now I need to get the md array setup for backups and stuff.  Wheeeeeeee.