Hi Falko, I have a very big problem at the moment. I followed your howto Backing Up And Restoring Your Dedicated Server With SystemImager. I also have a Strato root server but as it seems Strato has decided to move from BusyBox to LFS based rescue system. And there are some components missing. First of all there is no /etc/init.d/ directory in this new rescue system so having a look at this post I created one for storing the functions (right?). But then the script stops with this error Code: get_arch DISKORDER=sd,cciss,ida,rd,hd /tmp/backup.master: line 105: logmsg: command not found /tmp/backup.master: line 121: logmsg: command not found /tmp/backup.master: line 133: logmsg: command not found /tmp/backup.master: line 158: logmsg: command not found /tmp/backup.master: line 159: logmsg: command not found Disk geometry for /dev/hda: 0.000-58644.140 megabytes Disk label type: msdos Minor Start End Type Filesystem Flags 1 0.031 54.909 primary ext2 boot 2 54.910 1082.504 primary linux-swap 3 1082.505 58643.525 primary ext3 /tmp/backup.master: line 164: logmsg: command not found Killing off running processes. kill: usage: kill [-s sigspec | -n signum | -sigspec] [pid | job]... or kill -l [sigspec] udp-receiver: no process killed rsync: no process killed write_variables \s \r \m\ \n LFS based rescue system Unfortunately your suggestion from this post could not help me because now I simply get this: Code: get_arch DISKORDER=sd,cciss,ida,rd,hd Disk geometry for /dev/hda: 0.000-58644.140 megabytes Disk label type: msdos Minor Start End Type Filesystem Flags Killing off running processes. kill: usage: kill [-s sigspec | -n signum | -sigspec] [pid | job]... or kill -l [sigspec] udp-receiver: no process killed rsync: no process killed write_variables \s \r \m\ \n LFS based rescue system Your script is very important for me, since it seems to be the only way of restoring my test system in which I investigated a lot of time and the sollution could be interesting for many outher Strato users who like to use your script. I hope you can help me. Thanks in advance.
OK, I symlinked the logger of the rescue system with logmsg, so now I have logging to /var/log/messages enabled but this didn't help me any further. Here my steps so far: Code: cd /tmp/ wget http://www.falkotimme.com/howtos/dedicated_server_backup_restore_systemimager/restore chmod 755 restore ln -s /bin/logger /usr/bin/logmsg chmod 755 /usr/bin/logmsg mkdir /etc/init.d ./restore Output of /var/log/messages: Code: Jul 25 19:26:52 rescue root: enumerate_disks Jul 25 19:26:52 rescue root: hda Jul 25 19:26:52 rescue root: DISKS=1 Jul 25 19:26:52 rescue root: Partitioning /dev/hda... Jul 25 19:26:52 rescue root: Old partition table for /dev/hda: Jul 25 19:26:52 rescue root: parted -s -- /dev/hda mklabel msdos || shellout As far as I could see, the script still was't trying to restore the image and I didn't no why. It stops around here: Code: DISK_SIZE=`parted -s $DISK0 unit MB print | grep 'Disk geometry for' | sed 's/^.*-//g' | sed 's/\..*$//' | sed 's/MB//' ` [ -z $DISK_SIZE ] && shellout As I had a closer look to the parted params I found out that the version of the rescue system doesn't support "unit MB" for parted. So I simply deleted it because it uses MB as standard unit. Now it seems to work. At lease it is running... More later.
Thanks to Falko's work and his script I was able to restore my system but I'm a little bit afraid if it really worked completely as I get the following error messages at the end of the systemimager.master script. Does anybody know what that means and if it is critical? Code: receiving file list ... done ./ sent 109 bytes received 66 bytes 116.67 bytes/sec total size is 0 speedup is 0.00 You shouldn't call /sbin/grub-install. Please call /usr/sbin/grub-install instead! Probing devices to guess BIOS drives. This may take a long time. df: Warning: cannot read table of mounted file systems df: Warning: cannot read table of mounted file systems Could not find device for /boot: Not found or not a block device. Probing devices to guess BIOS drives. This may take a long time. Probing devices to guess BIOS drives. This may take a long time. /tmp/hXXXXXX.master: line 282: [INTERFACE0]: command not found /tmp/hXXXXXX.master: line 283: DEVICE: command not found /tmp/hXXXXXX.master: line 284: EOL: command not found umount /a/proc || shellout umount /a/boot || shellout umount /a/ || shellout Another problem is, that si_prepareclient doesn't include all installed kernel modules. I'm missing ipv6 although it's included in lsmod (see below) Code: Adding: /lib/modules/2.6.8-3-686/kernel/fs/mbcache.ko Adding: /lib/modules/2.6.8-3-686/kernel/fs/jbd/jbd.ko Adding: /lib/modules/2.6.8-3-686/kernel/fs/ext3/ext3.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/net/mii.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/net/e100.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/base/firmware_class.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/net/tg3.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/net/e1000/e1000.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/ide-core.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/ide-disk.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/pci/generic.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/pci/amd74xx.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/net/via-rhine.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/pci/piix.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/pci/via82cxxx.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/scsi/scsi_mod.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/scsi/sd_mod.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/scsi/libata.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/scsi/sata_svw.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/scsi/sata_nv.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/ide/ide-generic.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/md/md.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/md/raid1.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/video/cfbfillrect.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/video/cfbimgblt.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/video/cfbcopyarea.ko Adding: /lib/modules/2.6.8-3-686/initrd/vesafb.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/video/console/font.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/unix/unix.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/char/rtc.ko Adding: /lib/modules/2.6.8-3-686/kernel/security/commoncap.ko Adding: /lib/modules/2.6.8-3-686/boot/capability.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/md/dm-mod.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/md/dm-mirror.ko Adding: /lib/modules/2.6.8-3-686/kernel/drivers/md/dm-snapshot.ko Adding: /lib/modules/2.6.8-3-686/kernel/fs/ext2/ext2.ko Adding: /lib/modules/2.6.8-3-686/kernel/fs/quota_v1.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/packet/af_packet.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/ip_tables.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/iptable_filter.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/ip_conntrack.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/iptable_nat.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/iptable_mangle.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/ip_conntrack_ftp.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/ipt_LOG.ko Adding: /lib/modules/2.6.8-3-686/kernel/net/ipv4/netfilter/ipt_state.ko This produced some postfix errors about non supported protocols. Any ideas why it isn't included by default? This is the former output of lsmod on the old machine: Code: Module Size Used by ipt_state 2080 2 ipt_LOG 6560 0 ip_conntrack_ftp 72368 0 iptable_mangle 2880 0 iptable_nat 25156 0 ip_conntrack 35432 3 ipt_state,ip_conntrack_ftp,iptable_nat iptable_filter 2880 1 ip_tables 18496 5 ipt_state,ipt_LOG,iptable_mangle,iptable_nat,iptable_filter ipv6 265124 49 af_packet 22600 4 quota_v1 3552 2 ext2 71912 1 dm_snapshot 19100 0 dm_mirror 22580 2 dm_mod 59804 2 dm_snapshot,dm_mirror capability 4520 0 commoncap 7232 1 capability rtc 12760 0 unix 28788 363 font 8320 0 vesafb 6656 0 cfbcopyarea 3872 1 vesafb cfbimgblt 3040 1 vesafb cfbfillrect 3776 1 vesafb raid1 17856 0 md 49832 1 raid1 ide_generic 1408 0 sata_nv 8516 0 sata_svw 6820 0 libata 41700 2 sata_nv,sata_svw sd_mod 21728 0 scsi_mod 125228 2 libata,sd_mod via82cxxx 14332 0 piix 13440 1 via_rhine 21704 0 amd74xx 14620 0 generic 4416 0 ide_disk 19296 4 ide_core 139940 6 ide_generic,via82cxxx,piix,amd74xx,generic,ide_disk e1000 85924 0 tg3 74384 0 firmware_class 10176 1 tg3 e100 32608 0 mii 5120 2 via_rhine,e100 ext3 127432 1 jbd 62616 1 ext3 mbcache 9348 2 ext2,ext3 So on the new machine I'm missing ipv6 although I have a line alias net-pf-10 ipv6 in /etc/modprobe.d/aliases and there is a directory /lib/modules/2.6.8-3-686/kernel/net/ipv6 with content. To reenable ipv6 I had to run Code: modprobe ipv6 and to test if the module was loaded successfully Code: lsmod |grep -w 'ipv6' and finally to enable the automatic startup with Code: vi /etc/modules.conf alias net-pf-10 ipv6 But the module is still not loaded automatically. What am I missing? One last thing: The ispconfig user is only shown as 110 in ps -aux. Is this correct? Shouldn't it be something like admispconfig??
It seems like a few files are missing or truncated, e.g. /etc/modules (that's where the kernel modules are lsited that should be loaded automatically) and /etc/passwd (that's for the usernames). Did you restore the image onto the same hard disk, or was it maybe slightly smaller?
It's not the same hdd since Strato has replaced the server due to a memory fault. The hardware seems to be exactely the same but perhaps the disk is a little bit bigger or smaller. Is there anything I can do now?
