PART A OF THIS HOW-TO SHOWS YOU HOW TO UPGRADE REDHAT 7.3 TO 9 ON ANY SERVER.

PART B SHOWS YOU HOW TO UPGRADE CPANEL TO WORK WITH REDHAT 9

Last updated: 6th December 2003


This How-To is based on newuser's one here:
http://forum.rackshack.net/showthread.php?...?threadid=29540

and nuclear's one here:
http://forum.rackshack.net/showthread.php?...?threadid=24433

All credit goes to the both of them. My How-To uses the methods of both of them, and it is working brilliantly for myself. I thought I would share my hybrid method as well, since there is some interest from some members. I apologize if I am repeating a few things of what has already been said. It is noted that the above 2 How-To's do not describe all the necessary steps, which I will try to include here. My How-To therefore incorporates the 2 How-To's above, by including steps from both of them in a sequential manner that works, as well as documenting steps that were not mentioned in either of them, as well as mentioning potential errors that may arise and how to fix them.


Before you start, remember:

1) You need to BACK UP all data before attempting this, in case you do something irreversible.

2) You need to give yourself at least 4 hours just in case. If all goes well this should not take more than 2 hours. It took me 4 hours as I had to do things by trial and error.

3) If something goes wrong, DO NOT GIVE UP. I encountered at least 20 errors, and had to repeat certain steps at least 5 times to get it right. AS LONG AS YOU HAVE SSH ACCESS, YOU CAN FIX IT. Do not be too quick to order a restore - keep persisting if you encounter an error. I will try to document all the errors I came across.

4) If you are a newbie, don't try this. This is not for the novice, as errors WILL happen - it's a matter of fixing them.


At the end of this:
You should be running RH 9 with the latest version of cPanel.


Let's begin...


PART A: UPGRADING REDHAT FROM 7.3 TO 9 - THIS APPLIES TO ANY SERVER


1) Log in as root in SSH




2) Move your Apache log directory by running this command - we will be moving it back at the end (REMEMBER to move it back or apache won't restart):

mv /etc/httpd/logs /etc/httpd/logs2




3) Download and install apt - we will use this to upgrade from 7.3 to 9.0 - (alternatively, you can use up2date or yum, but apt worked for me - apt is a cool app ported from Debian) I found that it is possible to go straight from 7.3 to 9.0:

wget http://www.oktagone.net.au/files/apt-0.5.5...rh73.2.i386.rpm

rpm -i apt-0.5.5cnc5-fr0.rh73.2.i386.rpm




4) Edit the apt source list:

pico /etc/apt/sources.list

Comment out the 7.3 bits, and uncomment the 9 bits so that RH 9 will be installed. Save the file.




5) Upgrade using apt now by running the following commands:

apt-get update

apt-get dist-upgrade

You may need to run this to fix dependency problems as per instructions given to you by apt:

apt-get -f install

Only run it if you are told to do so.

If you get this error:

file /usr/bin/pear from install of php-4.2.2-17.2 conflicts with file from package php-devel-4.1.2-7.3.6
E: Sub-process /bin/rpm returned an error code (156)

Run this:

rpm -e php-devel-4.1.2-7.3.6

Then run this again:

apt-get -f install

It should now work. Once you have run it, run the first 2 commands again:

apt-get update

apt-get dist-upgrade

You may need to do the above a couple of time. When you get no errors, and start downloading between 100-250 Mb of RPMs for RH 9, that's when apt is upgrading your box. This can take 30-60 minutes.




6) Okay, your box now has Red Hat 9 on it. However, your kernel is still 7.3. So what do we do? We upgrade our kernel using up2date. At the same time, we make sure we have all the latest RPMs etc. for RH 9 from RHN as well.

You must first import the RHN security key with this command:

rpm --import /usr/share/rhn/RPM-GPG-KEY




7) Now, configure up2date to enable kernel upgrades by running this:

service named start

up2date --configure

Look for "pkgSkipList" and remove "kernel" from it by pressing on "C" to clear it so that kernel is NOT skipped.




Run the following commands to upgrade the kernel:

up2date -uv

up2date -p (Do not skip this command to update packages list, or else cPanel will not upgrade properly I found to my detriment)

up2date -uv

lilo -v -v

sync




9) Okay, time to reboot the box! Here's where you start praying.

reboot




10) Okay, that was the EASY bit. Your box now should be running RH 9 with the latest kernel. To check, log back in as root in SSH and run the following:

uname -r

2.4.20-24.9smp

cat /etc/redhat-release

Red Hat Linux release 9



CONGRATULATIONS - Your box now has Red Hat 9 installed! However, this is just the halfway mark.




PART B: UPGRADING CPANEL ON YOUR REDHAT 9 BOX




11) Time to upgrade cPanel, and recompile Apache. Here is where the headaches will begin. If you follow the following method, I hope you get less problems than I had to overcome. Run the following commands first:

ln -s /usr/lib/libpng.so.3 /usr/lib/libpng.so.2

rm /etc/httpd/logs

mv /etc/httpd/logs2 /etc/httpd/logs

You are basically fixing linbpng so that linuxconf/linuxhooks works, and moving the Apache log directly back to where it belongs. It is important to do these things first.




12) Upgrade cPanel. Using /scripts/upcp did not work for me, so it's best to use the following method instead:

wget layer1.cpanel.net/latest

sh latest

The above should start installing cPanel. This will also install a basic Apache, PHP setup. IT WILL NOT REPLACE YOUR CPANEL SETTINGS. So don't worry.

If the above starts installing cPanel 7.3 components instead of cPanel 9 components, run this again:

up2date -uv

up2date -p

up2date -uv

Then run step 12 again.

If that doesn't work, then try running this again:

apt-get update

apt-get dist-upgrade

Then run step 12 again.

If that still doesn't work, edit the redhat-release file and change:

Red Hat Linux release 9 (Shrike)

to:

Red Hat Linux release 9

In other words, remove (Shrike).

Then run step 12 again.

If that still doesn't work, then run this:

/scripts/distupgrade

Cpanel should then be upgraded properly.




13) Recompile Apache by running the following, and choosing your Apache and PHP modules:

/scripts/easyapache

Apache should restart automatically. If it doesn't start again, we are faced with an error.

a) libssl error

One error that I kept getting was this:

error while loading shared libraries: libssl.so.2: cannot open shared object file: No such file or directory

To fix this, I had to run apt again:

apt-get update

apt-get dist-upgrade

This will only reinstall mod_ssl again. Once this is done, try restarting apache first:

service httpd restart

If that didn't work, recompile apache:

/scripts/easyapache

If that still doesn't work, try running up2date:

up2date -uv

up2date -p

up2date -uv

That should fix it. If that still doesn't work, the you need to reinstall modssl manually by installing this file:

openssl096b-0.9.6b-6.i386.rpm


Another error:

b) httpd logs not found etc.

This should not happen if you delete the log directory and move the original log directory back.




14) Once cPanel and Apache are done, you can access your WHM. If you cannot access WHM in secure mode, but you are able to in unsecure mode, then you need to do this:

a) Run WHM in unsecure mod
b) Reset SSL for WHM/ cPanel by clicking the option in the left column
c) WHM should run now in secure mod again
d) Change your Root password just in case




15) IMAP may refuse to start for you. If this is the case, this is how to fix it:

edit /etc/xinetd.d/imap

Replace this line:

server = imapd

With this line:

server = /usr/sbin/imapd

Then restart IMAP:

service xinetd restart




16) EXIM may not restart at the end of all this. It was fine for me, but for some users it did not start up properly or was not installed. If this is the case for you, run this:

/scripts/exim4




Okay, that should be it. If I can remember anything else, I will post it here. I hope this helps someone else. The above did work for me.

Remember, as long as you have SSH access, NEVER GIVE UP. You can avoid a restore by persisting with whatever errors that come up.

I will keep updating this first post as other members post their tips and ways to fix their own errors in this thread.

Thanks for this How-To. I might try it when I have time. I guess it should also work for RH 7.2.

Thanks erwin. I used the up2date method to get the files and then used your comments here to finish the job. Everything is up and running, except that I cannot connect to get mail from my server either by pop nor imap.

I had the imap problem, and fixed it with your tip so now pop3 and imap services are both up and running. Cpanel and WHM are both running. The system reboots fine. RHN reports it as running RH9.

At first I thought the upgrade might have messed with my APF install. However, I stopped APF and tried to connect to the mail server with no luck.

So I'm definately confused here.

*Edit* Upon rebooting of the machine cpanel is not responding, though whm still is. Trying to troubleshoot that right now. Perhaps do a reinstall.

I think i'll wait untill I hear about any more bad experiences... BTW erwin, what is better about 9 compared with 7???

Well, one fact is that RH 7.x won't be "supported" anymore by the end of the year by Redhat. So no use of up2date anymore.

Well all cpanel took was a restart and now it's working fine. Still have the problem accessing email though, which is weird as all my account settings work just fine and were carried over, and pop3 and imap both start fine.

RH 9 has improve hyper-threading for Dual Xeons... but if you don't run applications that make use of it, then it makes no difference. EOL is only 4 months after RH 7.3, so although that's one reason to upgrade, it's only delays EOL by 4 months.

As for emails not working - keep trying to find out why. Emails work fine on my server. At least you are 99% done.

I ended up putting in a support ticket with cpanel and here it was either the wrong version of imap or it was pointing to the wrong install. They weren't too clear on it. But now everything is 100% working.. I didn't even have to touch my APF install which suprised me a bit.

Oh one last question... what does cpanel report your redhat as? Mine appears as:
[PHP] WHM 7.4.2 cPanel 7.4.2-S82
RedHat - WHM X v2.1
[/PHP]

Hmm, only 4 month...maybe then I let it be...

Good to hear it works.

Yeah, mine says that too.

If you run:

cat /etc/redhat-release

at the command line it should say RH 9 though.

During apt-get -f install at the end i get this


Need to get 0B/129MB of archives.
After unpacking 87.7MB of additional disk space will be used.
Do you want to continue? [Y/n] Y
Executing RPM (-Uvh)...
Preparing... ########################################### [100%]
file /usr/bin/pear from install of php-4.2.2-17.2 conflicts with file from package php-devel-4.1.2-7.3.6
E: Sub-process /bin/rpm returned an error code (156)





How can i fix this?

I had that problem and just removed the offending package like so:

rpm -e php-devel-4.1.2-7.3.6

And then I re-ran 'apt-get -f install' again.

That's not an official endorsement or anything, so if you screw your box it's not my fault. It seems to have worked for me though.

I added that fix to my first post, just in case it happens to anyone else.

Something else people might experience:

After all was said and done exim had totally vanished from the running processes on the server. Not only that but the startup script was missing from /etc/rc.d/init.d. I have no clue what the heck happened during my upgrade process that caused this, but it was an easy fix. I just ran /scripts/exim4 and exim was 'upgraded' and reappeared on the server... it even restarted on it's own.

I'll wait a week or two when I'm brave enough to try this

same thing happened to me.. i did the same thing you did and everything is ok.

HOWEVER.. cant get into the WHM for some reason..

Ok I got in through 2086.. however.. when i went to reset SSl cert.. i got this error.

Ehh now i get this error at

apt-get dist-upgrade


Need to get 0B/190MB of archives.
After unpacking 46.7MB of additional disk space will be used.
Do you want to continue? [Y/n] y
Executing RPM (-Uvh)...
Preparing... ########################################### [100%]
E: Sub-process /bin/rpm recieved a segmentation fault.


What do i do here?

I've already posted in the first post about this error - just read it.

I have no idea. Haven't seen that before.

Try running :

apt-get -f install

And see if that solves it.

Erwin,

Thanks for the great tip! I haven't yet put any client accounts on my new cPanel server, so I figured I'd give it a go. Piece o' cake!

There were a few errors, but pretty much just what you mentioned in your email.

One thing that may be a gotcha to some: when I did the part where you get the latest cpanel and then run it via sh, it did a completely new install and reset all settings to the defaults.

This was fine for me since I was still getting everything working the way I wanted, but may cause problems for others. When you run the latest cpanel installer, it gives you five (5) seconds to bail out of it if you don't want to run the installer.

I'm not sure what would happen to a cpanel installation that had lots of configurations and clients and such. I don't know if it would wipe them out or leave them as-is. Perhaps someone can comment on this.

EDIT: it looks like a good portion of my settings were in fact saved, so maybe even though it _looked_ like it was resetting everything, it didn't actually reset everything.

Anyway, for someone like me that just got the machine and is still setting it up, this was an excellent HOW-TO. I'm happy to be running the latest and greatest of everything now!

Danimal

Updating cPanel will NOT replace your current cPanel settings. I know, because I've done it. So your cPanel settings will NOT be reset to defaults, and should be save.

error while loading shared libraries: libssl.so.2 after upgrade and Internal server error

Had to rebuild apache . Thank you Erwin. You are a god

Glad to know it worked.

That error has already been documented in my first post since I had to fix it myself, and the fix provided.

Is layer1 the unstable release? Can you do the same with the stable release?




12) Upgrade cPanel. Using /scripts/upcp did not work for me, so it's best to use the following method instead:

wget layer1.cpanel.net/latest

sh latest

From the cPanel site:

http://www.cpanel.net/downloads.htm



Please select a download area:

I am curious as to how apt-rpm and up2date get along when run on the same system. Does apt-rpm simply act as a wrapper to rpms? In other words, does it just check for the installed rpms and update them, such that if a package was installed by up2date, apt-get update would still be able to update the programs without messing up up2date (and vice versa). Or does apt-rpm only update the programs that it installs?

All of these systems still use the rpm database. Apt, Yum, and up2date all just use their own system of looking for packages and resolving dependancies. As long as you use the default repository in the /etc/apt/sources.list file then you'll definately be fine as they mirror the redhat rpm server. It's only when you use multiple repositories that you MAY run into problems.

I install almost all my software on my home RH9 workstation using apt.

I get an error in trying to run the command in the howto:
up2date -uv

HERE IS THE ERROR
++++++++++++++++++
Error communicating with server. The message was:

Error Message:
Please run rhn_register (or up2date --register on Red Hat Linux 8.0)
as root on this client
Error Class Code: 9
Error Class Info: Invalid Server Certificate.
Explanation:
An error has occurred while processing your request. If this problem
persists please enter a bug report at bugzilla.redhat.com.
If you choose to submit the bug report, please be sure to include
details of what you were trying to do when this error occurred and
details on how to reproduce this problem.
++++++++++

Obviously I need to run the rhn_register again, but when I do, it says "this system appears to already be registered".

Do I need to re-register? Will I need to pay for the updates? I won't know the information that rackshack originally registered with.

well, it looks like I just lost $70 to a system restore!! I got to the reboot part and tried rebooting. My server never came back. I'm putting in a trouble ticket now... hopefully rackshack will fix it without having to do a restore.

Hey All,

Erwin I followed your instructions, rebooted with the new kernel and it came up fine. But when I run that command to find out which release I have it still says Redhat 7.3 ??

Please help!!!

TioChaharbaghi:

did you type:
uname -r

what kernel does it say? Is is the new kernel or the old one?

hey bobby, its the new one

it says kernel-2.4.20-20.9

Well, I'm back up and running. I don't know why my machine wasn't coming back after reboot. Rackshack fixed it and didn't tell me why.


Now that I'm back, I'm having problems with imap. Under Service Status its showing that there is a problem. I tried doing what the tutorial says, but there is no file called /etc/xinetd.d/imap for me to edit!!


TUTORIAL
+++++++++++
15) IMAP may refuse to start for you. If this is the case, this is how to fix it:

edit /etc/xinetd.d/imap

WOW!! everything is broken now!!!

Apache, Bind, Imap, and Horde are all showing problems in the service status. However, apache seems to be working because my websites are showing up. Horde won't work, and I assume its because Imap is down. I tried removing and re-installing imap, which didn't work. I checked the /etc/xinetd.d/imap file and the server=/usr/sbin/imap line is there.

apt-get dist-upgrade fails and when I type apt-get -f install, it wants to install httpd. However, there is already a httpd installed and the install exits with:
++++++
error: unpacking of archive failed on file /etc/httpd/logs: cpio: rename failed - Is a directory
E: Sub-process /bin/rpm returned an error code (1)
++++++

ARRGHHH!!!

Anyone that has lots of free time and would like to help me out, my email is:
bobbybobbertson@hotmail.com

thanks

Something that I noticed is that Cpanel still tries to download the upgrades from the Redhat 7.3 directory on Cpanel site.
Regards

You need to move the httpd log directory again.

Read my instructions carefully.

I have already posted the fix to most of the errors you will encounter. The service status in WHM always lag behind - if you can access WHM, it means that httpd is up and running.

I did say that this How-To is not for the beginner.

If that is happening, you need to run:

up2date -p

Then,

up2date -uv

Then upgrade cPAnel again. That happened to me too. I did post about the need to run up2date -p. If that didn't work, run apt again, then up2date.

All these errors you guys are reporting have occured to me, and I posted the relevant fixes. Like I said, ERRORS WILL HAPPEN. You need to fix them for your box.

I did put that directory back. But apt-get dist-upgrade must have screwed it up, when it tried to install http itself. Luckily I copied that directory, and didn't just move it, so when I copied it over again things were better.

However Imap is still failing and I have checked the /etc/xinetd.d/imap file and it contains the correct "server" reference.

Hi.

When I get to the end of step 5, I downloaded all the new pakeges, then I run

apt-get dist-upgrade it gives me

313 packages upgraded, 34 newly installed, 22 replaced, 0 removed and 0 not upgraded.
Need to get 0B/190MB of archives.
After unpacking 46.7MB of additional disk space will be used.
Do you want to continue? [Y/n] Y

I continue then this happens

Executing RPM (-Uvh)...
Preparing... ########################################### [100%]
E: Sub-process /bin/rpm recieved a segmentation fault..

And then the process stops. Any Ideas. Thanks

here is an answer from google

http://mail.linux.ie/pipermail/ilug/2003-J...une/004008.html

Thanks, I tried that with no luck. After that I tried

apt-get clean (cleared all my downloaded rpms)

and then

apt-get update

apt-get dist-upgrade

It now worked but when I get to the next steep

up2date -uv

I get this error

There was an SSL error: [('SSL routines', 'SSL3_GET_SERVER_CERTIFICATE', 'certificate verify failed')] A common cause of this error is the system time being incorrect. Verify that the time on this system is correct.

I cheeked the time and it’s correct. I looking on google now for some answers, has anybody run into this one.

Thanks

for that, I think you need to run rhn_register. it will say you are already registered, but do it agian. I had to.

I just tired and with same results, I will try to email redhat networks and see what they have to say.

I'm having the same problem as jwpexe. Please post an answer if you figure it out. Followed all the directions exactly, and got that error after trying to run 'up2date -uv'. I get the same error when I attempt to register the machine, or do anything else with the 'up2date' command.

If it's of any use or correlation, my server was just restored (new hard drive, fresh cPanel install) by RackShack (this is essentially the first work I've done to it since the server was returned to service).

This is from the Redhat Network homepage:




I think that probably explains the problems that other users and I are experiencing with up2date.

http://forum.rackshack.net/showthread.php?...&threadid=31347

Others too.

Maybe they should have put that in big letters and in a purple box on there home page to tell us about this, oh wait they did.

So I just downloaded the new up2date version an it works fine now.

The updates are at
https://rhn.redhat.com/errata/RHSA-2003-267...Hat%20Linux%209

Maybe. But after all the frustration it must be nice to know it wasnt the upgrade to 9