Can someone please confirm or laff at me for misunderstanding how i can nest installs please, my network at work is very locked down so i try to do the following.
On the edge dmz i have a full 1.17 install, I setup local repos for all the files needed to create an offline/airgapped install inside my network.
On my internal network, i create a new centos 7.4 install and setup repos to install from my edge server with no other dns.
I run the scenario katello installer but each time i get to about 82% when its trying to setup the cpdb database it fails because the logical expectation is wrong… so is this expected
I am able to do the following using Foreman1.16 and the above was not expected given my limited understanding
Morning,
Its at the poing the installer is populating the cpdb database at the 82% part of the script, You see it opening the db up and i think about to populate it and the logical test fails, i think it is expecting the open to fail because the db should not exist in a new install but instead of return 0 it returns 1 and then everything is red text
Sorry i can be of more use, i dont know how to use this pastebin thing
@fdisk101 installing a “disconnected” setup should be supported, if it bugs out for you it’s a bug that we should fix
can you provide the contents of /var/log/foreman-installer from the failing node? I wonder if that’s something like https://bugzilla.redhat.com/show_bug.cgi?id=1501846 (even tho that one should be already fixed in all relevant branches).
liquibase.exception.ChangeLogParseException: Error parsing line 7 column 75 of db/changelog/20171017062314-add-target-index-on-job-table.xml: schema_reference.4: Failed to read schema document 'http://www.liquibase.org/xml/ns/dbchangelog/dbchangelog-3.5.xsd', because 1) could not find the document; 2) the document could not be read; 3) the root element of the document is not <xsd:schema>.
which versions of liquibase and candlepin do you have on your install?
not sure, i built my upstream server and the failed downstream server from the foreman repos at the weekend, upstream is showing candlepin 2.1.12, liquidbase 3.1.0.1
Happy Mondays…
I set my upstream server to look at candlepin nightly as suggested, resynced, updated and the end result is i now have a internal hub server… woohoo, i had pissed away nearly a week of effort thinking it was me or my bloody office network. Many thanks, much appreciated help
it does, however, leave an annoying question unanswered… how come i can build a katello 3.6 build with a vm with internet connectivity using candlepin 2.1.12 as is… but when i do the disjoined version it fails because it needed the corrected 2.1.14 build.
Time to update my edge server
so the underlying problem is that: the candlepin RPM references a liquibase XSD file with a version, however, the version shipped in the liquibase RPM is lower, so the XML parser is nice to you and just goes and fetches the version off the internet. in your disconnected environment, the last step fails.
the fix is to require a lower version of that XSD file, which is actually shipped
This should now be fixed in Candlepin 2.1.14, newly packaged with Katello 3.5.1. The change was just published, so you can update using yum. Please let me know if it doesn’t work for you.
