Error upgrading 2024R1.4.3 to 2024R1.4.4

[eric.culpepper]

I'm attempting to update our Nagios XI 2024R1.4.3 to 2024R.1.4.4 deployed on Oracle Linux 8 through the admin interface and I'm receiving this error during the MySQL backup -

Backing up MySQL databases...
mysqldump: [Warning] Using a password on the command line interface can be insecure.
mysqldump: Got errno 28 on write
Error backing up MySQL database 'nagios' - check the password in this script!

[eric.culpepper]

As a guess, I checked the /store partition and found it only had ~3Gb of additional space. I cleared out old backups and the upgrade continued a bit further and it is now failing with this message -

Build process completed successfully
Installing '/usr/lib64/php/modules/ssh2.so'
install ok: channel://pecl.php.net/ssh2-1.4.1
configuration option "php_ini" is not set to php.ini location
You should add "extension=ssh2.so" to php.ini
Proceeding with upgrade...
Updating XI config wizards...
./upgrade: line 1416: cd: ./subcomponents/xiwizards: No such file or directory


I inherited this installation and I'm guessing it was a Nagios Core that was upgraded to Nagios XI at some point as there are /usr/local/nagios and /usr/local/nagiosxi directories. Any guidance would be appreciated. Previous updates through the admin web interface have worked without an issue in the 2024R1.x.x branch.

[eric.culpepper]

Any guidance on this issue? I searched in the /usr/local directory where Nagios is installed and do not appear to have an xiwizards directory matching the error message from the upgrade attempt:

$ sudo find /usr/local -type d -name subcomponents
/usr/local/nagiosxi/html/includes/components/xicore/images/subcomponents
$ sudo find /usr/local -type d -name xiwizards
$

[atc0005]

Eric,

It's not just you. I encountered the same error this AM when I attempted to update our Nagios XI instance from 2024R1.4.3 to 2024R1.4.4. Our Nagios XI host is RHEL 8 with the latest updates applied.

At this point rolling back (as I did) and then submitting a trouble ticket is probably the only viable next step.

I inherited this installation and I'm guessing it was a Nagios Core that was upgraded to Nagios XI at some point as there are /usr/local/nagios and /usr/local/nagiosxi directories.

FWIW, Nagios Core is bundled with Nagios XI.

[atc0005]

The 2024R1.4.3 to 2024R1.4.4 update completed successfully on an instance created from an OVA import (using nagiosxi-2024R1.4.3-OVA.zip).

The specific error we're encountering suggests that the upgrade script isn't finding the subcomponents/xiwizards subdirectory path within the decompressed xi-2024R1.4.4.tar.gz tarball.

Our maintenance window is long expired at this point, otherwise I'd attempt another in-place upgrade to see if maybe we had just encountered a temporary issue that led to a partially downloaded xi-2024R1.4.4.tar.gz tarball.

[atc0005]

I diffed the copy of the "upgrade" script between the 1.4.3 and 1.4.4 releases.

I suspect the problem is here (still digging):

Code: Select all

			#
			# Special handling for RHEL 8 systems to get libssh2 1.10.0
			# 
			if [ "$dist" == "el8" ]; then
				wget https://libssh2.org/download/libssh2-1.10.0.tar.gz
				tar -zxvf libssh2-1.10.0.tar.gz
				cd libssh2-1.10.0

				# Configure and install libssh2:
				./configure
				make
				make install
			else
				yum install -y libssh2 libssh2-devel
			fi

The CWD is changed and I don't see it being set back to what it needs to be before the XI Wizards update attempt.

Going to perform a fresh installation of Nagios XI 2024R1.4.3 on a new RHEL 8 instance and then attempt an update to 2024R1.4.4.

The OVA installation that I performed an in-place upgrade on uses CentOS 9 Stream, so it would not trigger that code block.

[atc0005]

Going to perform a fresh installation of Nagios XI 2024R1.4.3 on a new RHEL 8 instance and then attempt an update to 2024R1.4.4.

That was enough to reproduce the issue, both when using the web UI and when using the manual approach:

Code: Select all

cd /tmp
rm -rf nagiosxi xi*.tar.gz
wget http://assets.nagios.com/downloads/nagiosxi/xi-latest.tar.gz
tar xzf xi-latest.tar.gz
cd nagiosxi
./upgrade

If I repeat the upgrade attempt this forces the upgrade process to take another code path which avoided the problematic change of CWD from the libssh2 installation. That said, I can't speak to whether this omits anything important as I've not studied the upgrade logic closely enough. YMMV.

[atc0005]

Quick recap:

Once the first update attempt fails (assuming the libssh2 library successfully builds & installs), perform a follow up and it should complete successfully.

If you want to avoid the initial failure, this approach works (thanks to Nagios Support):

Code: Select all

cd /tmp
rm -rf nagiosxi xi*.tar.gz
wget http://assets.nagios.com/downloads/nagiosxi/xi-latest.tar.gz
tar xzf xi-latest.tar.gz
cd /tmp/nagiosxi
sed -i s/'make install'/"make install\n\t\t\t\tcd ..\n"/g  upgrade
./upgrade

This has worked without issue for me during testing. We'll upgrade our production instance during our next maintenance window. I don't anticipate any further issues with the update from 2024R1.4.3 to 2024R1.4.4 based on the testing performed thus far.

[eric.culpepper]

Thank you for your detailed help on this. I've managed to update the system but had a hiccup.

I followed the instructions for running sed across the upgrade script and it got further into it before failing with this error:
Build process completed successfully
Installing '/usr/lib64/php/modules/ssh2.so'
install ok: channel://pecl.php.net/ssh2-1.4.1
configuration option "php_ini" is not set to php.ini location
You should add "extension=ssh2.so" to php.ini
./upgrade: line 1336: nagiosxi/nagiosxi-db/mods/mysql/schema_70000.sql: No such file or directory


I tried running upgrade again and it finished on the second attempt. I had an issue with some python modules missing but I think those were customs installed by previous admins.

Thanks again, I appreciate the help!