Red Hat Bugzilla – Bug 1388201
candlepin content_id changes cause manfiest import/refresh failure
Last modified: 2018-10-29 03:58:24 EDT
Description of problem: Manifest upload on a second org failed Version-Release number of selected component (if applicable): # rpm -qa | grep satellite satellite-installer-6.3.0-1.el7sat.noarch satellite-cli-6.3.0-1.0.git.7.fb12bf2.el7sat.noarch tfm-rubygem-foreman_theme_satellite-1.0.0-1.git.2.94b76fc.el7.noarch satellite-6.3.0-1.0.git.7.fb12bf2.el7sat.noarch How reproducible: Always Steps to Reproduce: 1. Upload a manifest in org - `Default Organization` 2. Create a second manifest from the same portal account and upload to second org. Actual results: Manifest upload failed. See attached errors from: * production.log: 2016-10-24 12:50:04 [app] [E] Error during manifest import: {"displayMessage"=>"Failed to import archive", "requestUuid"=>"398feb37-f6bd-4709-998a-630f0e8f4a95"} 2016-10-24 12:50:04 [foreman-tasks/action] [E] Failed to import archive (Katello::Errors::CandlepinError) * /var/log/candlepin/error.log: 2016-10-24 12:50:04,662 [thread=http-bio-8443-exec-7] [req=398feb37-f6bd-4709-998a-630f0e8f4a95, org=secondorg] ERROR org.hibernate.engine.jdbc.spi.SqlExceptionHelper - ERROR: duplicate key value violates unique constraint "cp_content_label_key" Detail: Key (label)=(rhel-7-server-containers) already exists. 2016-10-24 12:50:04,693 [thread=http-bio-8443-exec-7] [req=398feb37-f6bd-4709-998a-630f0e8f4a95, org=secondorg] ERROR org.candlepin.sync.Importer - Failed to import archive Expected results: Manifest upload works successfully. Additional info:
Created attachment 1213534 [details] production.log
Created attachment 1213535 [details] candlepin error.log
Would we want to make sure users run a backup before modifying the DB as a "Step 0" or is it reasonably safe?
It appears that workaround mentioned in Comment 6 can cause breaking linkages. The candlepin team is working on a fix.
A safer alternative would be the following: Step 1: Remove the uniqueness constraint on the cp_content label column and change it to a regular index. # sudo su - postgres # psql postgres=# \connect candlepin candlepin=# alter table cp_content drop constraint cp_content_label_key; candlepin=# create index cp_content_label_key on cp_content (label); Step 2: upload the second org manifest again Step 3: refresh the first org manifest just to make sure it works.
Build : Satellite 6.2.4 Snap 1 rpm -qa | grep satellite tfm-rubygem-foreman_theme_satellite-0.1.33-1.el7sat.noarch satellite-cli-6.2.4-1.0.el7sat.noarch satellite-installer-6.2.0.12-1.el7sat.noarch satellite-6.2.4-1.0.el7sat.noarch Verification steps: 1. Upload a manifest in org - `Default Organization` 2. Created a second manifest from the same portal account and upload to second org - orgA. The 2 manifests for the 2 organizations were uploaded successfully Please see the screenshots
Created attachment 1216410 [details] Default org manifest upload
Created attachment 1216411 [details] OrgA manifest upload
follow up question: I've applied the workaround, but now the upgrade procedure does not work: Upgrade Step: migrate_candlepin... [ INFO 2016-11-03 14:04:29 verbose] Upgrade Step: migrate_candlepin... ########## ERROR ############ Error running command: liquibase --driver=org.postgresql.Driver --classpath=/usr/share/java/postgresql-jdbc.jar:/var/lib/tomcat/webapps/candlepin/WEB-INF/classes/ --changeLogFile=db/changelog/changelog-update.xml --url=jdbc:postgresql:candlepin --username=candlepin --password=UwAUXGvbHF3xAQcDeNsXSLah5bo5DoSb migrate -Dcommunity=False Status code: 65280 Command output: Liquibase update Failed: Migration failed for change set db/changelog/20161025100925-remove-unique-content-name-constraint.xml::20161025100925-1::awood: Reason: liquibase.exception.DatabaseException: Error executing SQL ALTER TABLE public.cp_content DROP CONSTRAINT cp_content_label_key: ERROR: constraint "cp_content_label_key" of relation "cp_content" does not exist Migrating candlepin database Traceback (most recent call last): File "/usr/share/candlepin/cpdb", line 245, in <module> dbsetup.update() File "/usr/share/candlepin/cpdb", line 69, in update self._run_liquibase("db/changelog/changelog-update.xml") File "/usr/share/candlepin/cpdb", line 92, in _run_liquibase self.community)) File "/usr/share/candlepin/cpdb", line 38, in run_command error_out(command, status, output) File "/usr/share/candlepin/cpdb", line 46, in error_out raise Exception("Error running command") Exception: Error running command [ERROR 2016-11-03 14:04:38 verbose] ########## ERROR ############ Error running command: liquibase --driver=org.postgresql.Driver --classpath=/usr/share/java/postgresql-jdbc.jar:/var/lib/tomcat/webapps/candlepin/WEB-INF/classes/ --changeLogFile=db/changelog/changelog-update.xml --url=jdbc:postgresql:candlepin --username=candlepin --password=UwAUXGvbHF3xAQcDeNsXSLah5bo5DoSb migrate -Dcommunity=False Status code: 65280 Command output: Liquibase update Failed: Migration failed for change set db/changelog/20161025100925-remove-unique-content-name-constraint.xml::20161025100925-1::awood: Reason: liquibase.exception.DatabaseException: Error executing SQL ALTER TABLE public.cp_content DROP CONSTRAINT cp_content_label_key: ERROR: constraint "cp_content_label_key" of relation "cp_content" does not exist Migrating candlepin database Traceback (most recent call last): File "/usr/share/candlepin/cpdb", line 245, in <module> dbsetup.update() File "/usr/share/candlepin/cpdb", line 69, in update self._run_liquibase("db/changelog/changelog-update.xml") File "/usr/share/candlepin/cpdb", line 92, in _run_liquibase self.community)) File "/usr/share/candlepin/cpdb", line 38, in run_command error_out(command, status, output) File "/usr/share/candlepin/cpdb", line 46, in error_out raise Exception("Error running command") Exception: Error running command Upgrade step migrate_candlepin failed. Check logs for more information. what is the recommended approach to resolve it?
Hi, Today the simplest option is to recreate the old constraint as the upgrade is failing because it recognizes the database is in an unknown state. If there is a way the installer could check whether the manual patch was taken then there is a manual liquibase task that could be used to mark the offending changeset as having already been run. http://www.liquibase.org/documentation/ant/marknextchangesetran_ant_task.html The manual SQL would be the reverse of the current alter table cp_content drop constraint cp_content_label_key; create unique index cp_content_label_key on cp_content (label);
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:2699