miq 14 to 15 dnf upgrade issue

[tjyang]

• Notify via twitter first about version 15 ;)
  • Read the miq 15 blog
• Look at upgrade page
• Need to figure out the repo full path by replacing 14/n* with 15/o*

rpm -ivh  https://rpm.manageiq.org/release/15-oparin/el8/noarch/manageiq-release-15.0-1.el8.noarch.rpm
[root@localhost vmdb]# dnf repolist |grep manage
manageiq-14-najdorf        ManageIQ 14 (Najdorf) Release- x86_64
manageiq-14-najdorf-noarch ManageIQ 14 (Najdorf) Release- noarch
manageiq-15-oparin         ManageIQ 15 (Oparin) Release - x86_64
manageiq-15-oparin-noarch  ManageIQ 15 (Oparin) Release - noarch
[root@localhost vmdb]#

• OS info

[root@localhost vmdb]# vmdb;cat VERSION;echo;cat /etc/redhat-release;date
najdorf-1.3
CentOS Stream release 8
Fri Jan  6 20:36:27 EST 2023
[root@localhost vmdb]#

• Run the dnf update and encounter following erros.

[root@localhost vmdb]# yum update
CentOS Stream 8 - AppStream                                                                                                               3.5 MB/s |  27 MB     00:07
CentOS Stream 8 - BaseOS                                                                                                                  6.2 MB/s |  26 MB     00:04
CentOS Stream 8 - Extras                                                                                                                   43 kB/s |  18 kB     00:00
CentOS Stream 8 - Extras common packages                                                                                                   12 kB/s | 5.2 kB     00:00
Ansible Runner for EL 8 - x86_64                                                                                                           25 kB/s | 8.3 kB     00:00
Extra Packages for Enterprise Linux 8 - x86_64                                                                                            4.6 MB/s |  13 MB     00:02
Extra Packages for Enterprise Linux Modular 8 - x86_64                                                                                    617 kB/s | 733 kB     00:01
Extra Packages for Enterprise Linux 8 - Next - x86_64                                                                                     105 kB/s | 179 kB     00:01
ManageIQ 14 (Najdorf) Release- x86_64                                                                                                     2.9 MB/s | 1.8 MB     00:00
ManageIQ 14 (Najdorf) Release- noarch                                                                                                      49 kB/s |  26 kB     00:00
ManageIQ 15 (Oparin) Release - x86_64                                                                                                     1.5 MB/s | 1.9 MB     00:01
ManageIQ 15 (Oparin) Release - noarch                                                                                                     343 kB/s | 242 kB     00:00
Error:
 Problem 1: package manageiq-gemset-15.1.1-1.el8.x86_64 requires libruby.so.3.0()(64bit), but none of the providers can be installed
  - cannot install the best update candidate for package manageiq-gemset-14.1.3-1.el8.x86_64
  - package ruby-libs-3.0.2-139.module_el8.5.0+916+17ac145a.x86_64 is filtered out by modular filtering
  - package ruby-libs-3.0.2-140.module_el8.6.0+989+38dab702.x86_64 is filtered out by modular filtering
<snipped>
  - package postgresql-server-13.7-2.module_el8.6.0+1152+d0162c0a.x86_64 is filtered out by modular filtering
(try to add '--allowerasing' to command line to replace conflicting packages or '--skip-broken' to skip uninstallable packages or '--nobest' to use not only best candidate packages)
[root@localhost vmdb]#

• end

[Fryguy]

@bdunne Can you comment here? This might be the ansible 6 issue, but I thought that would be handled automatically?

[bdunne]

Yes, we depend on Ruby 3.0 now, so the ruby:3.0 module needs to be enabled before the RPMs can be upgraded. Also, in v15 (oparin) we also changed the appliance PostgreSQL version from 10 to 13. @jrafanie was working through some upgrade instructions, but I don't know if they made it into the documentation yet.

[jrafanie]

@tjyang

You'll need to do a database backup before you upgrade as postgresql 13 is the new required database version. You can also do pg_upgrade if you're familiar with this but it's not covered here.

Make sure you have enough space to backup your database to a file.

Do a backup to a file like this in a location with enough space:

pg_dumpall -c --if-exists | gzip > vmdb.pg.gz

You'll need to clear the data directory since postgresql 10 is not compatible with postgresql 13. Ensure your appliance is fully backed up and the database backup file is not in the data directory.

systemctl stop postgresql
rm -rf $APPLIANCE_PG_DATA

After modifying the dnf repo information, backing up your database, and clearing the data directory:

dnf -y module reset postgresql
dnf -y module enable postgresql:13
dnf -y module reset ruby
dnf -y module enable ruby:3.0

I haven't tested the following step but it should upgrade to ruby 3.0 and postgresql 13 and remove ruby 2.7 and postgresql 10.

dnf -y upgrade --allowerasing
dnf -y autoremove

Check ruby -v and psql --version show 3.0 and 13.x.

Initialize a new data directory with pg 13:

appliance_console_cli --internal --standalone --password smartvm

Restore the db with pg 13:

gunzip < <your_backup_file>  | psql postgres

Run rake db:migrate and rake evm:automate:reset. Then you can restart the appliance or just systemctl start evmserverd

This assumes you're upgrading an appliance running both the application and database. If you're running just the application and not the database, you only need to upgrade the packages via the dnf commands. If you're running just a database, you'll need to do the dnf commands and database commands and leave the application steps to the first appliance connecting to this appliance.

[jrafanie]

If this works for you @tjyang, can you let us know so we can add it to the website?

[tjyang]

@jrafanie. Certainly I will report back here. I will do a miq14 all-in-one to miq15 test using VMWare VM.

[tjyang]

I tested @jrafanie's procedure and documented it in gdoc here.