Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1919075

Summary: The elasticsearch jobs fail with error no permissions for [indices:admin/aliases/get]
Product: OpenShift Container Platform Reporter: Saurabh Sadhale <ssadhale>
Component: LoggingAssignee: Jeff Cantrill <jcantril>
Status: CLOSED ERRATA QA Contact: Qiaoling Tang <qitang>
Severity: high Docs Contact:
Priority: high    
Version: 4.5CC: akamarathunga, amsingh, anisal, anli, aos-bugs, apaladug, david.karlsen, hkang, jcantril, mdeloren, oarribas, ocasalsa, periklis, qitang, sparpate, stwalter
Target Milestone: ---Flags: jcantril: needinfo-
jcantril: needinfo-
jcantril: needinfo-
Target Release: 4.5.z   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: logging-exploration
Fixed In Version: Doc Type: Bug Fix
Doc Text:
collapses the multiple policy cronjobs to a single job with multiple tasks it runs: delete rollover The reasoning is there is a potential race condition between the previous jobs which both rely upon a -write alias that may lead to false information. Additionally, ES does not have transactions or is ACID. By converting these into tasks we execute for management we: potentially free disk for ES to do additional work give a better chance for the rollover to be successful
Story Points: ---
Clone Of: Environment:
Last Closed: 2021-03-17 15:21:04 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:
Bug Depends On: 1916910, 1928772    
Bug Blocks:    

Description Saurabh Sadhale 2021-01-22 04:04:33 UTC
Description of problem:
The elasticsearch jobs rollover,delete for app infra audit fails with the following error: 

~~~
{"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}
Error while attemping to determine the active write alias: {"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}
~~~

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. Deploy Cluster Logging with EO image elasticsearch-operator.4.5.0-202012120433.p0 and cluster-logging-operator-v4.5.0-202012120433.p0
2.The jobs should fail after installation of Cluster Logging with the above described error. 
3.

Actual results:
Jobs get into error state

Expected results:
Jobs should be successful. 

Additional info:

Comment 2 Hui Kang 2021-01-26 15:05:30 UTC
@sau, could you provide the must-gather for the cluster https://github.com/openshift/cluster-logging-operator/tree/master/must-gather#usage? Thanks.

Comment 4 Anand Paladugu 2021-02-01 18:52:18 UTC
Hi

Any update on this BZ ?

Thanks

Anand

Comment 6 Anand Paladugu 2021-02-04 16:52:28 UTC
hmm. I dont know if there any limits on google drive.  Can you directly access the must gather from the case then ?

Comment 8 Anand Paladugu 2021-02-22 13:43:45 UTC
Hui  Kang,

Can you also check the case 02828704 that has the same symptoms. ? 

Thanks

Anand

Comment 9 Anand Paladugu 2021-02-23 16:54:25 UTC
@jcantril 

My customer seems to have upgraded to 4.6.17 and they are seeing this issue there too.  Do you know if the issue impacted 4.6.17 and are you planning to port the fix to 4.6.z ?

Thanks

Anand

Comment 12 Qiaoling Tang 2021-03-11 08:16:07 UTC
Verified with elasticsearch-operator.4.5.0-202103060503.p0

Comment 14 errata-xmlrpc 2021-03-17 15:21:04 UTC
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 (OpenShift Container Platform 4.5.35 extras update), 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-2021:0784

Comment 15 Anand Paladugu 2021-03-21 22:31:50 UTC
Hi

My customer seems to have upgraded to 4.6.17 and they are seeing this issue there too.  Do you know if the issue impacted 4.6.17 and are you planning to port the fix to 4.6.z ?

Thanks

Anand

Comment 16 akamarathunga 2021-03-31 02:35:05 UTC
We are seeing the error OCP 4.6.20 cluster with 
 - CLO version 4.6.0-202103202154.p0 
 - Elasticsearch operator version 4.6.0-202103130248.p0

{"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}

Error while attemping to determine the active write alias: {"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}

Comment 17 akamarathunga 2021-03-31 02:35:30 UTC
We are seeing the error OCP 4.6.20 cluster with 
 - CLO version 4.6.0-202103202154.p0 
 - Elasticsearch operator version 4.6.0-202103130248.p0

{"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}

Error while attemping to determine the active write alias: {"error":{"root_cause":[{"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"}],"type":"security_exception","reason":"no permissions for [indices:admin/aliases/get] and User [name=system:serviceaccount:openshift-logging:elasticsearch, roles=[admin_reader], requestedTenant=null]"},"status":403}

Comment 20 David J. M. Karlsen 2021-04-21 20:21:09 UTC
Same problem with clean install of:
cluster-logging.5.0.2-18                  Red Hat OpenShift Logging          5.0.2-18                                             Succeeded
elasticsearch-operator.5.0.2-18           OpenShift Elasticsearch Operator   5.0.2-18                                             Succeeded

[2021-04-21T19:52:00,904][ERROR][c.a.o.s.a.BackendRegistry] [elasticsearch-cdm-fzzbb5dp-1] Cannot authenticate user because admin user is not permitted to login via HTTP
[2021-04-21T19:52:01,005][INFO ][c.a.o.s.p.PrivilegesEvaluator] [elasticsearch-cdm-fzzbb5dp-1] No cluster-level perm match for User [name=system:serviceaccount:openshift-operators-redhat:elasticsearch-operator, roles=[admin_reader], requestedTenant=null] Resolved [aliases=[*], indices=[*], allIndices=[*], types=[*], originalRequested=[], remoteIndices=[]] [Action [indices:admin/template/get]] [RolesChecked [admin_user]]
[2021-04-21T19:52:01,005][INFO ][c.a.o.s.p.PrivilegesEvaluator] [elasticsearch-cdm-fzzbb5dp-1] No permissions for [indices:admin/template/get]

Comment 22 akamarathunga 2021-07-02 05:30:28 UTC
Cluster version - 4.6.21
CLO version - 4.6.0-202106181629
ESO version - 4.6.0-202106181629

The index rollover jobs use service account 'system:serviceaccount:openshift-logging:elasticsearch'. Therefore, I added it to 'sgconfig/roles_mapping.yml' as follows.

sg_role_admin:
  users:
    - 'CN=system.admin,OU=OpenShift,O=Logging'
    - 'system:serviceaccount:openshift-logging:elasticsearch'
  backendroles:
    - 'elasticsearch-operator'

Then, ran the 'es_seed_acl' to update search guard permissions.

This has to be done on all ES pods. These changes are not persistent. If the pod gets recreated, need to execute the above steps.

Comment 23 Periklis Tsirakidis 2021-07-02 13:08:04 UTC
Hi everyone,

you are addressing an issue for the 4.6.z on a closed BZ for 4.5.z. I accidentally happened to check this out. I am advising you to follow-up on the 4.6.z advisories for a similar BZ next time. This is the appropriate way to track down if and when something got fixed.

For example in your case:
- The appropriate BZ is: https://bugzilla.redhat.com/show_bug.cgi?id=1929688
- The fix is shipped with: 4.6.23
- The advisory for 4.6.23 telling you that is: https://errata.devel.redhat.com/advisory/70859