Bug 1272238
| Summary: | Setting a host build state to true facts are lost | ||
|---|---|---|---|
| Product: | Red Hat OpenStack | Reporter: | Dave Sullivan <dsulliva> |
| Component: | foreman | Assignee: | Mike Burns <mburns> |
| Status: | CLOSED EOL | QA Contact: | Shai Revivo <srevivo> |
| Severity: | high | Docs Contact: | |
| Priority: | high | ||
| Version: | Foreman (RHEL 6) | CC: | bkearney, dmaley, dsulliva, hbrock, jcoufal, johfulto, kbasil, lzap, mburns, morazi, ohadlevy, rhos-maint, scohen, srevivo |
| Target Milestone: | async | Keywords: | ZStream |
| Target Release: | Installer | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | foreman-1.6.0.44-7.el6ost | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2016-09-29 13:39:42 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
Adding bkearney as well. Bryan, any thoughts on this? deleting facts (and reports) on host build was done on purpose, Could you please provide information on why you would need it to stay when rebuilding a machine? (granted that when the machine is re provisioned, its out of date / irrelevant facts are purged anyway). per customer.... Current workflow is: 1. Boot servers into discovered_hosts 2. Host added to "hosts" table from discovered_hosts (build=0) and assigned to proper groups,hostname,ip,etc.. 3. Server build is executed "build=1" and foreman-hooks (after_build) are executed. If any hooks were relying on facts_hash (ie: checking for processorcount,processor0,memorytotal,manufactuer,modelname,etc...) those hooks will fail to find those entries. Since most facts collected during discovered_hosts is hardware related, we expected those facts to remain (as hardware isn't changing), at least until the provision process is finished (ie: before_provision hook is executed). Hope this helps? In latest version of Discovery (upstream) we actually prevent deletion of some facts (prefixed with "discovery_"): https://github.com/theforeman/foreman_discovery/blob/develop/app/services/foreman_discovery/host_converter.rb#L22-L25 You can use similar code as a workaround for now, maybe it makes sense to keep *all* discovered facts during provisioning. Or at least some (configurable by an option) if that breaks provisioning (network detection might fail perhaps). Ohad, if it makes sense we can keep them all: http://projects.theforeman.org/issues/12244 This is on the foreman server side, right? not the discovery image? Right it's foreman server side.
[root@gss-rhos-3 foreman]# grep -B 5 -A 5 setBuild app/models/host/managed.rb
# Called by build link in the list
# Build is set
# The boot link and autosign entry are created
# Any existing puppet certificates are deleted
# Any facts are discarded
def setBuild
self.build = true
self.save
errors.empty?
end
So looks like self.save calls a before_save hook that calls clear_data_on_build
/usr/share/foreman/app/models/host/managed.rb
# Custom hooks will be executed after_commit
after_commit :build_hooks
before_save :clear_data_on_build
...
def clear_data_on_build
return unless respond_to?(:old) && old && build? && !old.build?
clearFacts
clearReports
end
If we know on provision/re-provision the facts are regenerated/overwritten. Why can't we just comment out clearFacts, not sure the implications of this though.
Dave, I believe your fix would work as well (for all usage cases and not just discovered hosts) - would that be sufficient? This bug was accidentally moved from POST to MODIFIED via an error in automation, please see mmccune with any questions Closing list of bugs for RHEL OSP Installer since its support cycle has already ended [0]. If there is some bug closed by mistake, feel free to re-open. For new deployments, please, use RHOSP director (starting with version 7). -- Jaromir Coufal -- Sr. Product Manager -- Red Hat OpenStack Platform [0] https://access.redhat.com/support/policy/updates/openstack/platform |
Description of problem: First make sure a host has fact values [root@gss-rhos-3 api_scripts]# psql -h localhost -U foreman -d foreman -c "select * from hosts where id=95" Password for user foreman: id | name | ip | last_compile | last_freshcheck | last_report | updated_at | source_file_id | created_at | mac | root_pass | serial | puppet_status | domain_id | architecture_id | operatingsystem_id | environment_id | subnet_id | ptable_id | medium_id | build | comment | disk | ins talled_at | model_id | hostgroup_id | owner_id | owner_type | enabled | puppet_ca_proxy_id | managed | use_image | image_file | uuid | compute_resource_id | puppet_proxy_id | certname | image_id | organization_id | location_id | type | otp | realm_id | compute_profile_id | provision_method | primary_interface ----+------------------------------------------+--------------+---------------------+-----------------+---------------------+----------------------------+----------------+----------------------------+----------- --------+------------------------------------+--------+---------------+-----------+-----------------+--------------------+----------------+-----------+-----------+-----------+-------+---------+------+----------- -----------------+----------+--------------+----------+------------+---------+--------------------+---------+-----------+------------+------+---------------------+-----------------+------------------------------ ------------+----------+-----------------+-------------+---------------+-----+----------+--------------------+------------------+------------------- 95 | rhosp-6.openstack.gsslab.rdu2.redhat.com | 10.10.72.104 | 2015-10-15 19:36:47 | | 2015-10-15 19:36:18 | 2015-10-15 19:37:03.029948 | | 2015-07-17 07:54:54.842362 | 40:f2:e9:8 5:82:82 | $1$T8aNAgKO$Sn.KLDP2hpD/rDow3QKzD. | | 0 | 1 | 1 | 1 | 1 | 1 | 12 | 7 | f | | | 2015-07-17 09:32:45.798942 | 2 | 19 | 3 | User | t | 1 | t | | | | | 1 | rhosp-6.openstack.gsslab.rdu2 .redhat.com | | | | Host::Managed | | | | build | eth2 (1 row) [root@gss-rhos-3 api_scripts]# psql -h localhost -U foreman -d foreman -c "select * from fact_values where host_id=95"Password for user foreman: id | value | fact_name_id | host_id | updat ed_at | created_at -------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--------------+---------+-------------- --------------+---------------------------- 17706 | be:fd:fe:81:b1:96 | 378 | 95 | 2015-08-11 19 :37:50.518444 | 2015-08-11 19:37:50.518444 17707 | be:fd:fe:81:b1:96 | 379 | 95 | 2015-08-11 19 :37:50.525294 | 2015-08-11 19:37:50.525294 17708 | 96:0c:48:88:3e:cb | 380 | 95 | 2015-08-11 19 :37:50.52995 | 2015-08-11 19:37:50.52995 turn on debugging. Edit ~foreman/config/environments/production.rb and uncomment this line: #apparently this is needed to reload foreman nicely touch ~foreman/tmp/restart.txt then I tail /var/log/foreman/production.log Then switch build to true/1 #!/bin/sh curl -s -H "Accept:application/json" -X PUT -u admin:vq8bD5S9DJamGaJa \ --cacert /var/lib/puppet/ssl/certs/gss-rhos-3.openstack.gsslab.rdu2.redhat.com.pem --insecure \ : -d "host[build]=1" https://gss-rhos-3.openstack.gsslab.rdu2.redhat.com/api/hosts/rhosp-6.openstack.gsslab.rdu2.redhat.com [root@gss-rhos-3 api_scripts]# sh addbuildfact_rhosp-6.sh Rerun query on fact_values [root@gss-rhos-3 api_scripts]# psql -h localhost -U foreman -d foreman -c "select * from fact_values where host_id=95 > " Password for user foreman: id | value | fact_name_id | host_id | updated_at | created_at ----+-------+--------------+---------+------------+------------ (0 rows) [root@gss-rhos-3 api_scripts]# So they have been wiped The log files show SQL (1.5ms) DELETE FROM "fact_values" WHERE (host_id = 95) SQL (24.7ms) DELETE FROM "reports" WHERE (host_id = 95) So why delete the fact values, might be another problem with deleting host from reports too. Now further problems noted on facts query as you see the error where we try to reference WHERE (hosts.id = 'rhosp-6.openstack.gsslab.rdu2.redhat.com') where that is really the hosts.name. [root@gss-rhos-3 api_scripts]# cat gethostfacts.sh #!/bin/sh curl -X GET -u admin:vq8bD5S9DJamGaJa \ -H "Accept:application/json" \ --cacert /var/lib/puppet/ssl/certs/gss-rhos-3.openstack.gsslab.rdu2.redhat.com.pem --insecure \ https://gss-rhos-3.openstack.gsslab.rdu2.redhat.com/api/v2/hosts/rhosp-6.openstack.gsslab.rdu2.redhat.com/facts FactValue Load (0.9ms) SELECT fact_values.id FROM "fact_values" INNER JOIN "hosts" ON "hosts"."id" = "fact_values"."host_id" AND "hosts"."type" IN ('Host::Managed') WHERE (hosts.name = 'rhosp-6.openstack.gsslab.rdu2.redhat.com') FactValue Load (0.5ms) SELECT fact_values.id FROM "fact_values" INNER JOIN "hosts" ON "hosts"."id" = "fact_values"."host_id" AND "hosts"."type" IN ('Host::Managed') WHERE (hosts.id = 'rhosp-6.openstack.gsslab.rdu2.redhat.com') PG::InvalidTextRepresentation: ERROR: invalid input syntax for integer: "rhosp-6.openstack.gsslab.rdu2.redhat.com" LINE 1: ...ts"."type" IN ('Host::Managed') WHERE (hosts.id = 'rhosp-6.o... ^ : SELECT fact_values.id FROM "fact_values" INNER JOIN "hosts" ON "hosts"."id" = "fact_values"."host_id" AND "hosts"."type" IN ('Host::Managed') WHERE (hosts.id = 'rhosp-6.openstack.gsslab.rdu2.redhat.com') SQL (38.4ms) SELECT "fact_values"."id" AS t0_r0, "fact_values"."value" AS t0_r1, "fact_values"."fact_name_id" AS t0_r2, "fact_values"."host_id" AS t0_r3, "fact_values"."updated_at" AS t0_r4, "fact_values"."created_at" AS t0_r5, "fact_names"."id" AS t1_r0, "fact_names"."name" AS t1_r1, "fact_names"."updated_at" AS t1_r2, "fact_names"."created_at" AS t1_r3, "fact_names"."compose" AS t1_r4, "fact_names"."short_name" AS t1_r5, "fact_names"."type" AS t1_r6, "fact_names"."ancestry" AS t1_r7, "hosts"."id" AS t2_r0, "hosts"."name" AS t2_r1, "hosts"."ip" AS t2_r2, "hosts"."last_compile" AS t2_r3, "hosts"."last_freshcheck" AS t2_r4, "hosts"."last_report" AS t2_r5, "hosts"."updated_at" AS t2_r6, "hosts"."source_file_id" AS t2_r7, "hosts"."created_at" AS t2_r8, "hosts"."mac" AS t2_r9, "hosts"."root_pass" AS t2_r10, "hosts"."serial" AS t2_r11, "hosts"."puppet_status" AS t2_r12, "hosts"."domain_id" AS t2_r13, "hosts"."architecture_id" AS t2_r14, "hosts"."operatingsystem_id" AS t2_r15, "hosts"."environment_id" AS t2_r16, "hosts"."subnet_id" AS t2_r17, "hosts"."ptable_id" AS t2_r18, "hosts"."medium_id" AS t2_r19, "hosts"."build" AS t2_r20, "hosts"."comment" AS t2_r21, "hosts"."disk" AS t2_r22, "hosts"."installed_at" AS t2_r23, "hosts"."model_id" AS t2_r24, "hosts"."hostgroup_id" AS t2_r25, "hosts"."owner_id" AS t2_r26, "hosts"."owner_type" AS t2_r27, "hosts"."enabled" AS t2_r28, "hosts"."puppet_ca_proxy_id" AS t2_r29, "hosts"."managed" AS t2_r30, "hosts"."use_image" AS t2_r31, "hosts"."image_file" AS t2_r32, "hosts"."uuid" AS t2_r33, "hosts"."compute_resource_id" AS t2_r34, "hosts"."puppet_proxy_id" AS t2_r35, "hosts"."certname" AS t2_r36, "hosts"."image_id" AS t2_r37, "hosts"."organization_id" AS t2_r38, "hosts"."location_id" AS t2_r39, "hosts"."type" AS t2_r40, "hosts"."otp" AS t2_r41, "hosts"."realm_id" AS t2_r42, "hosts"."compute_profile_id" AS t2_r43, "hosts"."provision_method" AS t2_r44, "hosts"."primary_interface" AS t2_r45 FROM "fact_values" LEFT OUTER JOIN "fact_names" ON "fact_names"."id" = "fact_values"."fact_name_id" LEFT OUTER JOIN "hosts" ON "hosts"."id" = "fact_values"."host_id" AND "hosts"."type" IN ('Host::Managed') WHERE (fact_names.name <> '_timestamp') ORDER BY "fact_values"."value" ASC NULLS FIRST LIMIT 9999 OFFSET 0 Rendered api/v2/fact_values/index.json.rabl within api/v2/layouts/index_layout (3.3ms) (1.8ms) SELECT COUNT(DISTINCT "fact_values"."id") FROM "fact_values" LEFT OUTER JOIN "fact_names" ON "fact_names"."id" = "fact_values"."fact_name_id" WHERE (fact_names.name <> '_timestamp') Body: { "total": 1685, "subtotal": 10, "page": 1, "per_page": 9999, "search": " host = rhosp-6.openstack.gsslab.rdu2.redhat.com", "sort": { "by": null, "order": null }, "results": {"rhosp-2.openstack.gsslab.rdu2.redhat.com":{"swapfree_mb":"0.00","swapfree":"0.00 MB","boardproductname":"00FG986","bios_release_date":"05/14/2014","uniqueid":"0a0a6548","network_eth2":"10.10.72.0","gluster_host_ip":"10.10.72.101","ipaddress_eth2":"10.10.72.101","ipa_host_ip":"10.10.72.101","network_br_ex":"10.10.77.0","ipaddress_br_ex":"10.10.77.107","vlans":"101,102,100","swapsize_mb":"1024.00","swapsize":"1024.00 MB","augeasversion":"1.1.0","network_lo":"127.0.0.0","ipaddress_lo":"127.0.0.1","iptables_version":"1.4.21","ip6tables_version":"1.4.21","memorysize":"15.49 GB","memorytotal":"15.49 GB","memorysize_mb":"158 The results return are actually bogus, not sure why we'd return results back for fact_values based on timestamp anyway. In customers environment no results are returned. Version-Release number of selected component (if applicable): My testing was done on [root@gss-rhos-3 api_scripts]# rpm -qa | grep foreman foreman-installer-1.6.0-0.4.RC1.el7ost.noarch foreman-discovery-image-7.0-20150227.0.el7ost.noarch foreman-proxy-1.6.0.30-6.el7ost.noarch rubygem-foreman_api-0.1.11-6.el7sat.noarch foreman-postgresql-1.6.0.49-6.el7ost.noarch foreman-selinux-1.6.0.14-1.el7sat.noarch ruby193-rubygem-foreman_discovery-1.3.0-2.3.el7ost.noarch rubygem-hammer_cli_foreman-0.1.1-16.el7sat.noarch ruby193-rubygem-foreman_openstack_simplify-0.0.6-8.el7ost.noarch openstack-foreman-installer-3.0.24-1.el7ost.noarch ruby193-rubygem-foreman-tasks-0.6.9-1.1.1.el7ost.noarch foreman-1.6.0.49-6.el7ost.noarch Customer's environment is more important though in regards to needing a fix first. # rpm -qa |grep -i foreman rubygem-hammer_cli_foreman-0.1.1-16.el6sat.noarch ruby193-rubygem-foreman_column_view-0.2.0-1.el6.noarch ruby193-rubygem-foreman_openstack_simplify-0.0.6-8.el6ost.noarch rubygem-hammer_cli_foreman-doc-0.1.1-16.el6sat.noarch ruby193-rubygem-foreman_hooks-0.3.8-1.el6.noarch ruby193-rubygem-foreman_column_view-doc-0.2.0-1.el6.noarch bac-RHEL6-ost5-foreman-1.0-1.noarch foreman-1.6.0.44-6.el6ost.noarch ruby193-rubygem-foreman_discovery-1.3.0-2.2.el6ost.noarch foreman-discovery-image-7.0-20140905.0.2.el7sat.noarch openstack-foreman-installer-2.0.34-1.el6ost.noarch foreman-proxy-1.6.0.33-2.el6ost.noarch ruby193-rubygem-foreman_hooks-doc-0.3.8-1.el6.noarch foreman-installer-1.6.0-0.2.RC1.el6ost.noarch foreman-mysql2-1.6.0.44-6.el6ost.noarch foreman-selinux-1.6.0.14-1.el6sat.noarch How reproducible: See above Actual results: facts still be available after setting hosts build to 1/true Expected results: Additional info: Also concerned about hosts being deleted from reports table