I’m running CentOs4 with kernel 2.6.9-89.EL currently.
There is a security notification:
with regards a security flaw which my hosting provider (RackSpace) is enforcing on all servers.
However, I’m trying to update using yum update but the latest kernel build is not available.
2.6.9-89.0.9.EL or above is secure from this vulnerability.
Anyone got a method to get up to this version on VM Pro?
It should show up in a “yum update” – if not, you can also manually pulling it down with something like this:
yum install kernel-2.6.9-89.0.9.EL
That’s all provided by the standard CentOS respository – if the above doesn’t work, you may want to verify that there’s nothing wrong with your CentOS repository config in /etc/yum.repos.d/.
Thnks, after buggering around in the repos. yum was doing lots of segmentation faults on doing a full update.
Now yum says it has installed the latest kernel 2.6.9-89.0.11.EL but there is no initrd for it and grub.conf hasn’t been updated. Keeps hanging. I think I have shafted it.
Are you sure you have enough space on /boot?
You can determine that by running “df -h”.
Also, is there enough RAM on your system? Sometimes, yum can segfault in low-memory situations.
something major got broken and I have now got a freshly rebuilt RHEL server to get working again.
I don’t suppose anyone knows the best steps for disaster recovery?
- restore password files
- Install Virtualmin Pro
- copy /etc/webmin over the top
- copy /home directories
- copy /var/lib/mysql
Do you have the standard Virtualmin backups?
If so, restoring is as simple as:
Setup the OS
Install Virtualmin with the install.sh script
Go into Backups and Restore -> Restore Backup, and point it to the directory containing all your Virtualmin backups. From there, it’ll restore everything for you, including settings, users, home dir files, databases, and the like.
I have a complete file backup of the server though it’s just that a full system restore is not working so I’m left with doing it the hard way!
That’s a tough one then
Yeah, something along the lines of what you suggested should get you there as best as possible… of note though, is that databases generally can’t be reliably backed up using a typical filesystem backup, so you may see some database problems.
Just to complete the thread, I managed to get everything back and running in about 4 hours although I could have done it in 2 with hindsight and performing more operations simultaneously.
Here’s what I had:
New RH ES4 Server
Old drive mounted as /mnt/olddrive
Here’s what I did:
first I copied passwd*, shadow* and group* from the old drive into /etc
don’t forget to use -p when copying to maintain permissions.
then I copied over the /home folder (took a while),
stopped mysqld and copied /var/lib/mysql.
and up2date -if kernel
download and run VMPro install.sh (it stopped halfway through with no warning so I ran it again and it completed without issue)
yum upgrade php*
yum install php5*
copied over /etc/webmin and /usr/libexec/webmin
had a problem with the SSL cert in FF and had to regenerate it by using IE to access webmin.
copied * .cf and virtual * to /etc/postfix
copied over /etc/httpd/conf and conf.d
I had a problem with mysqld not starting and after checking the log it seemed I shouldn’t have copied the ib_logfiles over.
The final issue I had was that some of the sites weren’t coming up due to not being able to connect to mysql. It turned out these were php5 sites and that the /etc/php5.ini file wasn’t pointing to the php5 modules folder where the mysql.so was located (phpinfo() helped me figure that one out).