Web Hosting Talk







View Full Version : How to Restore off a mounted slave disk (crash/hack/bad kernel)


Steven
02-01-2004, 05:05 PM
How to Restore off a mounted slave disk on Redhat + Cpanel (crash/hack/bad kernel)

Warning: This is not to be done by the unexperienced admins bad things could happen. Hire someone.

Sometimes if you get hacked or boot with a bad kernel you need to have the datacenter mount your old drive, and install a new one with a fresh installation of redhat. First things first lets make sure you have an updated kernel, we will do it with up2date for ease of use:

up2date -f kernel

check the grub.conf /or lilo config to ensure the settings are right. (if using lilo run this command aswell /sbin/lilo -v -v and check for errors)

Now reboot the server


shutdown -r now




I. Mount the backup Drive

First check to see if there are any drives mounted.



df -h



You should get something simular to this if it isent mounted yet:


[root@localhost root]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/hda2 37G 1.4G 33G 4% /
/dev/hda1 101M 7.7M 88M 9% /boot
none 125M 0 125M 0% /dev/shm
[root@localhost root]#


if it is mounted you will see something like this:


-bash-2.05b# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/hda3 53G 31G 20G 62% /
/dev/hda1 99M 14M 80M 15% /boot
none 248M 0 248M 0% /dev/shm
/dev/hdc3 53G 31G 20G 61% /mnt/old
-bash-2.05b#


as you can see hdc3 is the old drive, and it is mounted as /mnt/old. Yours will be diffent possibly. But keep in mind you need to know the mount point for the backup drive for the following steps.

Lets mount if not already mounted.


fdisk -l


check for additional drives, in this tutorial we will use /dev/hdc3.


mount /dev/hdb3 /mnt/old


now that the drive is mounted you can browse files like:


ls /mnt/old/home



II. Rsync the files

Do these commands one by one,


rsync -vrplogDtH /mnt/old/usr/local/apache/conf /usr/local/apache
rsync -vrplogDtH /mnt/old/var/named /var
rsync -vrplogDtH /mnt/old/home/* /home
rsync -vrplogDtH /mnt/old/usr/local/cpanel /usr/local
rsync -vrplogDtH /mnt/old/var/lib/mysql /var/lib
rsync -vrplogDtH /mnt/old/var/cpanel /var
rsync -vrplogDtH /mnt/old/usr/share/ssl /usr/share
rsync -vrplogDtH /mnt/old/var/ssl /var
rsync -vrplogDtH /mnt/old/usr/local/cpanel/3rdparty/mailman /usr/local/cpanel/3rdparty
rsync -vrplogDtH /mnt/old/var/log/bandwidth /var/log
rsync -vrplogDtH /mnt/old/usr/local/frontpage /usr/local
rsync -vrplogDtH /mnt/old/var/spool/cron /var/spool


Next Step - Get some important files from etc:

cd /mnt/old/etc
rsync -vrplogDtH secondarymx domainalias valiases vfilters exim* proftpd* pure-ftpd* passwd* group* *domain* *named* wwwacct.conf cpupdate.conf quota.conf shadow* *rndc* ips* ipaddrpool* ssl /etc


Now would be a good time to change your password for root:


passwd


III. Updating software and restarting services


/scripts/upcp
/scripts/updatenow
/scripts/sysup
/scripts/fixeverything
/scripts/exim4
/scripts/easyapache
/scripts/securetmp


This is needed to update cpanel information. Please note, sshd might fail and not start after running fixeverything. You have to login to whm, and go to the rpm installer and "FORCE" install opensshd-server, opensshd, opensshd-client and then restart sshd from whm.

IV. Restarting services


/scripts/restartsrv httpd
/scripts/restartsrv cpanel
/scripts/restartsrv mysql
/scripts/restartsrv named
/scripts/restartsrv exim




After you are pleased that everything is working fine, restart the entire server.

This tutorial works 100% on redhat, 7.3, 8.0, 9.0 and fedora. it has been tested on those boxes. But remember, if you dont know what your doing them hire someone to check your server out.

scooterh
02-01-2004, 05:07 PM
Were you the one that wrote the same/similiar one over at EV1 forums?

Steven
02-01-2004, 06:30 PM
I used knowledge from that one and my own knowledge to make a more complete one.

sprintserve
06-13-2004, 10:07 AM
It will be good if you can credit the references if you can.

spho
07-26-2004, 05:39 PM
i followed what you said, but i cant find the folder etc on my old HDD can you please help ?

Techie411
09-19-2004, 11:53 PM
Hey Linuxguy....Do you know if this works with Free BSD??

Steven
09-20-2004, 12:22 AM
It should work just fine.

KingAdmin
09-20-2004, 02:07 PM
Originally posted by thelinuxguy
It should work just fine.

It will work fine on FreeBSD except slight changes in 2 lines:
rsync -vrplogDtH /mnt/old/var/lib/mysql /var/lib
should be:
rsync -vrplogDtH /mnt/old/var/db/mysql /var/db
and also after running "cd /mnt/old/etc" :
rsync -vrplogDtH secondarymx domainalias valiases vfilters exim* proftpd* pure-ftpd* passwd* group* *domain* *named* wwwacct.conf cpupdate.conf quota.conf shadow* *rndc* ips* ipaddrpool* ssl /etc
should be:
rsync -vrplogDtH secondarymx domainalias valiases vfilters exim* cpupdate.conf quota.conf master.passwd* *rndc* ips* ipaddrpool* ssl /etc

Other than that rest will work fine and should restore all accounts and settings.

mali
01-18-2005, 08:33 AM
Thank you for this valuable information.
I want to migrate Windows Plesk 7 from one machine to anothermachine.
Can you guide me how can i achieve this.

darken
05-20-2005, 05:22 PM
help!
I have followed the instructios for my server- who runs centOS, and I am getting an error message when trying to restart Cpanel "sorry I don't know about Cpanel". Can anyone please help?

Rockbottom
07-06-2005, 10:55 AM
I had a kernel panick on my server and got my datacenter to slave the hard drive and do an OS reload on a new drive. I am running CentOS 3.5. Will this tutorial be suitable for recovering all data and accounts etc?

fancy claps
07-13-2005, 06:17 PM
rockbottom, i just did what you'll be doing (on centos 3.5) and it worked fine. i didnt follow this exactly, but it should work alright.

the only problems i ran into were permissions issues on named, mysql, and the public_html folders, but those are easily solved and basic problems.

mousepotatoe
04-04-2008, 03:57 PM
I just used this step by step howto to recover from disaster and it was amazing! The only snag i ran into was time constraint when i ran
up2date -f kernel
That ran for a couple of hours before i killed it off and continued, but I achieved 100% success on:
WHM 11.15.0 cPanel 11.18.3-R21703 CENTOS Enterprise 4.6 i686 on standard - WHM X v3.1.0

Thank you for such a handy guide!

linux-tech
04-07-2008, 10:55 AM
Keep in mind that this thread was originally written 4 years ago, and that CPanel has changed a LOT since then. This will get you the starter, but with CP11 there came a lot more files to address, some will automatically fix themselves, some won't.

this thread (http://forums.cpanel.net/showthread.php?t=71394) will give you a bit more up to date restore procedures.

I'm not saying that Stevens thread is bad, but it is years out of date.

Vladimir Micovic
05-14-2009, 07:28 AM
great tutorial
5 cent for this!