I've encountered several upgrade errors at customers or redhatters when DHCP and/or DNS configuration simply disappears after Satellite upgrade. When our users add or modify subnets or domains, our installer overwrite the configuration with the initial one. Since our installer is limited and only supports definition of one subnet/domain, all the others are deleted and users need to restore those configurations from backups. Affected files are: /etc/zones.conf and /etc/dhcp/dhcpd.conf Add these steps during the upgrade (before katello-installer is executed): A) Our installer currently only supports one domain or subnet. Ask users to check those files manually before running the upgrade. B) If there is more than one entry (subnet or domain), they need to back up these files. C) After installer (--upgrade option) run, ask users to compare those files and put the manual changes back restarting affected services.
In addition to the items mentioned above, the user should be instructed to use the --capsule-dhcp-managed & --capsule-dns-managed switches if they are modifying the installer provided configuration files. --capsule-dhcp-managed DHCP is managed by Foreman proxy (default: true) --capsule-dns-managed DNS is managed by Foreman proxy (default: true) Also, we should document how the user can run the following to check for locally modified content before running the installer. katello-installer --noop --verbose --upgrade
Re-assigning to Stephen, who is currently working on upgrade-related content, after discussion with Ali. Stephen - would you be able to take a look at the changes requested above and see if they still hold true for the current upgrade process, and if this is covered in any of the other feedback we have received?
(In reply to Rich Jerrido from comment #1) > In addition to the items mentioned above, the user should be instructed to > use the --capsule-dhcp-managed & --capsule-dns-managed switches if they are > modifying the installer provided configuration files. > > --capsule-dhcp-managed DHCP is managed by Foreman proxy (default: > true) > --capsule-dns-managed DNS is managed by Foreman proxy (default: > true) > > Also, we should document how the user can run the following to check for > locally modified content before running the installer. > > katello-installer --noop --verbose --upgrade Hello Rich We need to say see logs, same as 6.1? Has the log name changed? 6.1 Docs says "Review the /var/log/katello-installer/katello-installer.log to see what changes would be applied if the --noop option was omitted." I just ran the command on my Sat6.2 test VM and it says "The full log is at /var/log/foreman-installer/satellite.log" Even without the --verbose option there is a lot there. What exactly should we look for? Thank you
(In reply to Stephen Wadeley from comment #9) > (In reply to Rich Jerrido from comment #1) > > In addition to the items mentioned above, the user should be instructed to > > use the --capsule-dhcp-managed & --capsule-dns-managed switches if they are > > modifying the installer provided configuration files. > > > > --capsule-dhcp-managed DHCP is managed by Foreman proxy (default: > > true) > > --capsule-dns-managed DNS is managed by Foreman proxy (default: > > true) > > > > Also, we should document how the user can run the following to check for > > locally modified content before running the installer. > > > > katello-installer --noop --verbose --upgrade > > Hello Rich > > We need to say see logs, same as 6.1? Has the log name changed? 6.1 Docs > says "Review the /var/log/katello-installer/katello-installer.log to see > what changes would be applied if the --noop option was omitted." > > I just ran the command on my Sat6.2 test VM and it says "The full log is at > /var/log/foreman-installer/satellite.log" > > Even without the --verbose option there is a lot there. What exactly should > we look for? > > Thank you You would be looking for files that are changed, similar to a UNIX diff (so lines that start with '---' and '+++'. Example follows: [DEBUG 2016-07-21 16:14:12 main] Executing 'diff -u /etc/default/pulp_workers /tmp/puppet-file20160721-6625-1uakf2r' [ WARN 2016-07-21 16:14:12 main] /Stage[main]/Pulp::Config/File[/etc/default/pulp_workers]/content: [ INFO 2016-07-21 16:14:12 main] --- /etc/default/pulp_workers 2016-07-21 16:13:09.844000000 -0400 [ INFO 2016-07-21 16:14:12 main] +++ /tmp/puppet-file20160721-6625-1uakf2r 2016-07-21 16:14:12.889000000 -0400 [ INFO 2016-07-21 16:14:12 main] @@ -2,7 +2,7 @@ [ INFO 2016-07-21 16:14:12 main] [ INFO 2016-07-21 16:14:12 main] # Define the number of worker nodes you wish to have here. This defaults to the number of processors [ INFO 2016-07-21 16:14:12 main] # that are detected on the system if left commented here. [ INFO 2016-07-21 16:14:12 main] -PULP_CONCURRENCY=2 [ INFO 2016-07-21 16:14:12 main] +PULP_CONCURRENCY=4 [ INFO 2016-07-21 16:14:12 main]
Thank you Rich, for comment 11
Hello I think the work for *this* bug is done. The entire upgrade procedure will be reviewed as part of the work on BZ#1339417. Thank you
This is now live on the customer portal. Satellite 6.2 docs are here: https://access.redhat.com/documentation/en/red-hat-satellite/