FATAL: kernel too old

M

Manie

Verified User
Joined
Jul 18, 2006
Messages
121
Location
Utrecht / Netherlands
I cleaned up an old running machine, which I offcourse, recently disabled in the backup cronjob because of high server load.

The machine hasn't beed updated in ages, and this week I updated some items in directadmin.

Now I received an e-mail stating:

Code: 
A new message or response with subject:

DirectAdmin has been updated

has arrived for you to view.
Follow this link to view it:

http://xxx.xxx.xxx.xxx:2222/CMD_TICKET?action=view&number=000001525&type=ticket


======================================================
Automatically generated email produced by DirectAdmin 1.33.7

Do Not Reply.


Well great, however I also received this e-mail right away!

Code: 
A new message or response with subject:

The service 'directadmin' on server xxx is currently down

has arrived for you to view.
Follow this link to view it:

http://xxx.xxx.xxx.xxx:2222/CMD_TICKET?action=view&number=000001526&type=ticket

======================================================
Automatically generated email produced by DirectAdmin 1.33.7

Do Not Reply.

This seems quite bad so I logged in and found out:

Code: 
virtual1:/home/admin# /etc/init.d/directadmin start
Starting DirectAdmin: FATAL: kernel too old
/etc/init.d/directadmin: line 20: 14346 Segmentation fault      $PROGBIN
[ FAILED ]

I tried upgrading the old Etch to the new Lenny, but I just have to much errors that I don't dear to reboot without a proper copy.

I also can not enter the controlpanel since DA won't start, and I don't have everything in up-to-date backups.

What is the best backup strategy for this if I am going to do a clean install of this machine? I can copy over all the /home directories, but that won't give me the users I need to restore DA.

I am open for help/suggestions!
Cause as soon kernel errors come in, my expertise is way over :-(

Kind regards,
Armand

P.S. This is becoming hell of a week for me, the second reinstall in 3 days :'-(
 
Need some insights on restoring:

1. If I do a fresh install, and I have the following type of admin_backups:
a. admin.root.admin.tar.gz
b. reseller.admin.reseller1.tar.gz
c. reseller.admin.reseller2.tar.gz
d. user.admin.user1.tar.gz
e. user.reseller1.user2.tar.gz

Should I restore them in that order as well?
Admin first, resellers second, users third?
 
If you update Debian v4 to v5, you need to update your DA license to v5 too, so email John using the safe submit form asking for an upgrade..... (if you haven't already)
 
Last edited:
Code: 
Backup all users
echo "action=backup&local_path=/home/admin/admin_backups&owner=admin&type=admin&value=multiple&when=now&where=local&who=all" >> /usr/local/directadmin/data/task.queue

/usr/local/directadmin/dataskq &
 
Manie said:
Should I restore them in that order as well?
Admin first, resellers second, users third?
Click to expand...

When I transfer all admin/resellers/users, I simply backup them all and restore them all at the same time, and it always works out.
 
I am sorry Floyd, but since directadmin is not a running process, this is not working.

Code: 
virtual1:/home/admin/admin_backups# /usr/local/directadmin/dataskq &
[2] 28387
[1]   Exit 127                Verified User /usr/local/directadmin/dataskq
 
So my plan is now to restore from the backupfiles I have, afterwards I copy the contents of all /home dir's to the new/fresh install.

That means copying for each user from full system copy:
/home/user/domains/
/home/user/imap/
/home/user/Maildir/
/home/user/public_html/

The databases are stored in:
/home/mysql
So I copy all the sitenaam_site files from the ons sysbackup to the new /home/mysql.

Question: will this work if I switch to mysql 5.1 in the fresh install? Or is it obviouse going to fail? (Tips: suggestions how to prevent, entire mysql dump better?)

Afterwards I probably need these commands to fix DA recognizing all the users:
http://help.directadmin.com/item.php?id=265
(That is if I missed a couple with the tar.gz and need to recreate them manually)

And same with:
http://help.directadmin.com/item.php?id=205 (proftpd passwords)

And this one to rebuild virtual:
http://help.directadmin.com/item.php?id=66 (If missing backups)

Am I missing something here?

Regards,
Armand
 
Manie said:
I am sorry Floyd, but since directadmin is not a running process, this is not working.
Click to expand...

DirectAdmin does not have to be running. However if DirectAdmin is really broken then you are right its not going to run the backups.

Maybe you can get a DirectAdmin binary from somebody that will work with Lenny.

Another option which is a lot harder is to open one of the backup files and see what DirectAdmin backs up and do it manually.
 
You are probably right Floid, however since I have a lot of users running on this system (200 domains) I am going the slow way.... moving around all the users to other machines :-(
 
Since I am still in the backup progress, I wish to share some nifty tricks to make this process a lot easier! :-)

I got old backups which I can restore, but need to make up-to-date with the current machine where DA isn't working anymore.
( Problem is I need to reallocate a lot of IP-addresses.)

My friend in the moving process:
( YES: I temporary enable ssh permit root login. )

/usr/bin/rsync --verbose --rsh=ssh --times --owner --group --links --perms --recursive --delete --force --exclude-from=/etc/rsync.exclude --stats /home/username/ [email protected]:/home/username/

This way Rsync matches the uid's and gid's on the target.

Offcourse I also need a quick up-to-date database for each user.

On source:
mysqldump -u da_admin -pMYPASSWORD web_site >> web_site.sql

Move to target:
scp -r web* [email protected]:/home/admin/restore/

On target:
mysql -u da_admin -pMYOTHERPASSWORD web_site < web_site.sql

Most of you would probably know all of this, but just wanted to point it out to people restoring from older backups.
 
You do realize that there is a lot more to backup than just /home and the databases.

/etc/virtual/
/etc/shadow
/etc/passwd
/etc/group
/etc/gshadow
/usr/local/directadmin/data/
/var/spool/virtual/ (if it exists)


Those are from memory. You should untar one of the old backups to see what needs to be backed up.
 
I do, but since the backups are two weeks old, I restored THOSE first.
But since my customers are kinda picky in losing anything, I am copying their userfiles (Maildir's) and Databases manually.

I fully understand that; to restore DA backups from cratch, isn't doable.
As soon as I see Maildir's missing I am aware that I need to fish in /etc/virtual as well.

I offcourse do not wish to suggest people that Rsync is the only thing you need to do.
However with changing uid's en gid's on a new system, AFTER you did a tar.gz restore, it is a very usefull tool!
 
Manie said:
I do, but since the backups are two weeks old, I restored THOSE first.
Click to expand...

Previously you had said:

Manie said:
I got old backups which I can restore, but need to make up-to-date with the current machine where DA isn't working anymore.
Click to expand...

I did not understand that to mean that you had actually restored the old backups first. I thought you were doing this INSTEAD of restoring the old backups.

DA Support probably could have helped you with this, perhaps even giving you a binary that would work with your system.
 
English can be very tricky. Most Americans cannot speak it well.
 
You must log in or register to reply here.
Back
Top