Description of problem: 1. After connecting with Openshift provider, 2. Navigating to compute --> Containers --> Pods then clicking on the filter 'namespace / Container group' 3. browser returns following "Unexpected error encountered Errors in Management Engine can be caused by: ① Accessing Management Engine from multiple tabs or windows of the same browser on a single machine. Close any duplicate browser sessions, then select a menu option above. ② Pressing the back button during a session. Close any duplicate browser sessions, then select a menu option above. ③ An internal system error. Please contact your administrator for assistance. Error text: PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist LINE 1: ...WHERE ((container_groups.deleted_on IS NULL) AND ("container... ^ : SELECT COUNT(DISTINCT "container_groups"."id") FROM "container_groups" LEFT OUTER JOIN "container_projects" ON "container_projects"."id" = "container_groups"."container_project_id" LEFT OUTER JOIN "ext_management_systems" ON "ext_management_systems"."id" = "container_groups"."ems_id" WHERE ((container_groups.deleted_on IS NULL) AND ("container_groups"."namespace" = 'default')) [container_group/show_list]" 4. Checked log and internal db schema, there is no namespace column on container_groups table. 5. Customer lab and production is having same issue. Version-Release number of selected component (if applicable): CFME 5.6.2.1(Customer lab env) CFME 5.6.0.13(Customer production env) OpenShift version 3.3 How reproducible: 1. Connect Openshift provider 2. Navigating to compute --> Containers --> Pods then clicking on the filter 'namespace / Container group' Steps to Reproduce: 1. 2. 3. Actual results: Returns unexpected error to browser with "PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist" Expected results: we should see info based on filter Additional info:
Log from UI appliance ################################# [tachoi@tachoi log]$ grep -b2 -i "PG::UndefinedColumn" production.log 284099687-[----] D, [2016-10-19T15:14:00.827768 #5368:15cd57c] DEBUG -- : MiqUserRole Inst Including Associations (0.3ms - 1rows) 284099809-[----] D, [2016-10-19T15:14:00.835352 #5368:15cd57c] DEBUG -- : (1.4ms) SELECT COUNT(DISTINCT "container_groups"."id") FROM "container_groups" LEFT OUTER JOIN "container_projects" ON "container_projects"."id" = "container_groups"."container_project_id" LEFT OUTER JOIN "ext_management_systems" ON "ext_management_systems"."id" = "container_groups"."ems_id" WHERE ((container_groups.deleted_on IS NULL) AND ("container_groups"."namespace" = 'default')) 284100265:[----] F, [2016-10-19T15:14:00.836074 #5368:15cd57c] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist 284100454-LINE 1: ...WHERE ((container_groups.deleted_on IS NULL) AND ("container...
Log 1. PG::UndefinedColumn: ERROR: column container_services.namespace does not exist LINE 1 #Checked lab db table structure but couldn't find namespace column from container_services ################################## vmdb_production=# select * from container_services \x\g\x Expanded display is on. -[ RECORD 1 ]---------------+------------------------------------- id | 1 ems_ref | cafd4331-6e0a-11e6-8e26-00163e36e921 name | docker-registry ems_created_on | 2016-08-29 17:05:31 resource_version | 2119 session_affinity | ClientIP portal_ip | 172.30.149.174 ems_id | 6 container_project_id | 1 service_type | ClusterIP container_image_registry_id | created_on | 2016-10-13 00:04:27.227478 Log 2 FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist #Checked lab db table structure but couldn't find namespace column from container_groups ################################## vmdb_production=# select * from container_groups \x\g\x Expanded display is on. -[ RECORD 1 ]------------+------------------------------------------------------------------ id | 1 ems_ref | ab6485bf-90d6-11e6-afff-00163e36e921 name | docker-registry-1-cax1g ems_created_on | 2016-10-12 23:50:35 resource_version | 3529 restart_policy | Always dns_policy | ClusterFirst ems_id | 6 container_node_id | 2 last_perf_capture_on | container_replicator_id | 1 ipaddress | type | ManageIQ::Providers::Kubernetes::ContainerManager::ContainerGroup container_project_id | 1 phase | Pending message | reason | container_build_pod_id | created_on | 2016-10-13 00:04:26.371143 deleted_on | old_ems_id | old_container_project_id |
Hi Gregg I haven't got update on this case, can you please provide update on this case? Regards, Taeho
Hi Ari Do you have any update on this issue? Regards, Taeho
Hi, Ive seen this bug before but no one managed to define whats supposed to happen when you press that filter button. I think the real solution is just to remove it entirely. Any objections?
Upstream PR: https://github.com/ManageIQ/manageiq/pull/12302
(In reply to Ari Zellner from comment #8) > Hi, Ive seen this bug before but no one managed to define whats supposed to > happen when you press that filter button. I think the real solution is just > to remove it entirely. > > Any objections? Yes I don't think it's useful as it is now (We can remove it). BTW by removing it what happens in old installations? Is this still saved in their DB (and it won't disappear)?.
(In reply to Federico Simoncelli from comment #10) > (In reply to Ari Zellner from comment #8) > > Hi, Ive seen this bug before but no one managed to define whats supposed to > > happen when you press that filter button. I think the real solution is just > > to remove it entirely. > > > > Any objections? > > Yes I don't think it's useful as it is now (We can remove it). BTW by > removing it what happens in old installations? Is this still saved in their > DB (and it won't disappear)?. Yeah, I think youre right since I just remove it from the fixtures. Whats the protocol in this case?
(In reply to Ari Zellner from comment #11) > Yeah, I think youre right since I just remove it from the fixtures. Whats > the protocol in this case? Check with Gregg T or Greg B. I am concerned with the fact that this will keep failing going forward for old installations. I think we'll have to find a solution to avoid support requests or duplicated bugs, etc.
Hi Team customer tried the patch in comment9 but still seeing same issue. no difference has been made. [tachoi@tachoi log]$ grep "MIQ(EvmApplication.start) EVM Startup initiated" evm.log [----] I, [2016-11-01T14:44:55.974259 #13405:ae5998] INFO -- : MIQ(EvmApplication.start) EVM Startup initiated [tachoi@tachoi log]$ grep "column container_groups.namespace does" production.log [----] F, [2016-11-01T13:50:04.250028 #3243:1270064] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T14:51:53.876260 #13711:12b9cb4] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T14:52:02.539753 #13703:109ce90] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T14:53:58.157242 #13703:109d3f4] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T14:59:22.690978 #13711:12b43cc] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T15:01:02.688449 #13703:109ce90] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T15:45:31.484294 #13703:109d1c4] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist [----] F, [2016-11-01T15:46:02.699063 #13711:12b9cb4] FATAL -- : Error caught: [ActiveRecord::StatementInvalid] PG::UndefinedColumn: ERROR: column container_groups.namespace does not exist Regards, Taeho
Hi Team Can I have update on this request? Customer is asking for progress on this issue Regards, Taeho
Right now it looks like we can prevent this from happening on new installations but not solve the issue on existing installations. Was your customer relying on this button? we plan to remove it.
New commit detected on ManageIQ/manageiq/euwe: https://github.com/ManageIQ/manageiq/commit/bf7399b31f956f26c03aae1c0eda63d52ada2c8f commit bf7399b31f956f26c03aae1c0eda63d52ada2c8f Author: Oleg Barenboim <chessbyte> AuthorDate: Thu Nov 3 17:40:26 2016 -0400 Commit: Oleg Barenboim <chessbyte> CommitDate: Mon Nov 7 15:53:05 2016 -0500 Merge pull request #12302 from zeari/remove_filters Remove filters from container `show_list` pages (cherry picked from commit d77f49cf860eae796645bff42f743c3257ab6188) https://bugzilla.redhat.com/show_bug.cgi?id=1386473 db/fixtures/miq_searches.yml | 30 ------------------------------ 1 file changed, 30 deletions(-)
Moving to Post. Tae, the best I can offer is a script that will fix this on existing installations. It would be a simple process. What it would do is delete the entry from MiqSearches so that filter button wont be available any more.
Hi Ari This is the answer from customer. "No plans on using that button, I think it can be removed" Please let me know, once the removal script is ready. Regards, Taeho
Hi Team Any update on the removing script ? Taeho
Hi Gregg, Thank you for your reply. Please allow me to double check the final workaround for existing CFME. 1. Remove the related records in miq_searches. >delete from miq_searches where name in ('default_Namespace / ContainerGroup', 'default_Namespace / ContainerService'); 2. Update the db/fixtures/miq_searches.yml according to https://github.com/ManageIQ/manageiq/pull/12302/files. Best Regards, Chen
I also cant find the old filter. I think the bug could have been on an instance that was upgraded and an old filter remained. either way, these filters were removed and so i dont think theres a problem here.
As I understand it this error happened on pressing a button which is now removed[1]. If the button is no longer available (and other similar buttons in other pages do not cause the bug) we can treat this as verified. [1] https://bugzilla.redhat.com/show_bug.cgi?id=1386473#c10
Verified. filter button no longer exists.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days