Bug 1638814 - cpu sockets for hypervisors are not available in Satellite
Summary: cpu sockets for hypervisors are not available in Satellite
Keywords:
Status: CLOSED DUPLICATE of bug 1647762
Alias: None
Product: Red Hat Satellite
Classification: Red Hat
Component: Subscription Management
Version: 6.3.3
Hardware: x86_64
OS: Linux
medium
medium
Target Milestone: Unspecified
Assignee: satellite6-bugs
QA Contact: Katello QA List
URL:
Whiteboard:
: 1638812 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2018-10-12 13:50 UTC by Kenny Tordeurs
Modified: 2019-08-12 19:34 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of: 1638812
Environment:
Last Closed: 2018-11-26 18:33:04 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Kenny Tordeurs 2018-10-12 13:50:35 UTC
Description of problem:
virt-who is sending the cpu sockets:
~~~
2018-10-12 15:18:34,395 [virtwho.main INFO] MainProcess(1535):Thread-2 @virt.py:_send_data:922 - Report for config "vcenter" gathered, placing in datastore
2018-10-12 15:18:36,552 [virtwho.main INFO] MainProcess(1535):Thread-3 @virt.py:_send_data:922 - Report for config "satotest" gathered, placing in datastore
{
    "hypervisors": [
        {
            "facts": {
                "cpu.cpu_socket(s)": "2",
                "hypervisor.type": "VMware ESXi",
                "hypervisor.version": "6.0.0"
            },
~~~

However no sockets are being reported in Satellite:
# curl -s -u $user:$password -k https://$satellite/api/v2/hosts/$host_id/| python -m json.tool | grep "socket(s)"

Sockets count for hypervisor should be included:
https://bugzilla.redhat.com/show_bug.cgi?id=1307024

Version-Release number of selected component (if applicable):
satellite-6.3.3-2.el7sat.noarch

How reproducible:
100%

Steps to Reproduce:
1. Setup virt-who to report hypervisors
2. Verify cpu socket count in Satellite
3.

Actual results:
No cpu socket count 

Expected results:
see cpu socket count

Additional info:

Comment 1 Brad Buckingham 2018-10-12 14:59:43 UTC
*** Bug 1638812 has been marked as a duplicate of this bug. ***

Comment 5 Kenny Tordeurs 2018-10-15 09:03:49 UTC
I looked a bit deeper into this and I believe it's caused because of the following:

#######
CPU count is available in candlepin:

[root@provisioning ~]# su - postgres -c "psql candlepin -c \"select * from cp_consumer where name='vmvisor1.gsslab.brq2.redhat.com' ;"\"""
                id                |          created           |          updated           | autoheal | entitlementstatus |              name               | servicelevel |   username    |                 uuid 
                | environment_id | consumer_idcert_id | keypair_id |             owner_id             | type_id | releasever |                       compliancestatushash                       |        lastchecki
n         | annotations | cont_acc_cert_id | content_access_mode | entitlement_count | recipient_owner_key 
----------------------------------+----------------------------+----------------------------+----------+-------------------+---------------------------------+--------------+---------------+----------------------
----------------+----------------+--------------------+------------+----------------------------------+---------+------------+------------------------------------------------------------------+------------------
----------+-------------+------------------+---------------------+-------------------+---------------------
 4028ef816668485c016668604f170004 | 2018-10-12 15:03:54.135+02 | 2018-10-12 15:09:45.699+02 | t        | valid             | vmvisor1.gsslab.brq2.redhat.com | Self-Support | foreman_admin | 2a0779f1-3735-4335-ad
f2-846da90a0e8e |                |                    |            | 4028ef8153198e100153198f8d470001 | 1004    |            | 4f2568fb098d97cd0089cff1ddf3f5f1aaddae6194b541a7b89662b9c7ca9ecc | 2018-10-12 15:05:
29.341+02 |             |                  |                     |                 0 | 
(1 row)

[root@provisioning ~]# su - postgres -c "psql candlepin -c \"select * from cp_consumer_facts where cp_consumer_id='4028ef816668485c016668604f170004';"\"""
          cp_consumer_id          |   element   |       mapkey       
----------------------------------+-------------+--------------------
 4028ef816668485c016668604f170004 | VMware ESXi | hypervisor.type
 4028ef816668485c016668604f170004 | 2           | cpu.cpu_socket(s)
 4028ef816668485c016668604f170004 | 6.0.0       | hypervisor.version
(3 rows)


#######

The cpu count that is shown in the webui and API call is coming from foreman db:

# su - postgres -c "psql foreman -c \"select * from fact_names where name like '%cpu_socket(s)%';"\"""
~~~
256 | cpu::cpu_socket(s)                 | 2016-07-30 13:38:27.649622 | 2016-07-30 13:38:27.649622 | f       | cpu_socket(s)      | Katello::RhsmFactName | 200
266 | lscpu::socket(s)                   | 2016-07-30 13:38:27.958328 | 2016-07-30 13:38:27.958328 | f       | socket(s)          | Katello::RhsmFactName | 165
~~~

Example for some host:
select value from fact_values where host_id=239 and (fact_name_id=256 OR fact_name_id=266);
~~~
 id   | value | fact_name_id | host_id |         updated_at         |         created_at         
-------+-------+--------------+---------+----------------------------+----------------------------
 19111 | 6     |          256 |     239 | 2017-01-12 13:34:23.36352  | 2017-01-10 14:22:02.557855
 19108 | 6     |          266 |     239 | 2017-01-12 13:34:23.362563 | 2017-01-10 14:22:02.552852
~~~

Also confirmed with API call:

curl -s -u $user:$password -k https://$satellite/api/v2/hosts/$host_id/| python -m json.tool | grep "socket(s)"
~~~
Sockets for  239 :
        "cpu::cpu_socket(s)": "6",
        "lscpu::socket(s)": "6",
~~~

#########

So for hypervisors the foreman db does not contain any facts about the CPU count so it's not available via API call or in the webui

~~~
Content Host Properties
- OS
- Architecture
- Number of CPUs
- Sockets
- Cores per Socket
- RAM (GB) => NaN
- Virtual Guest => No

This can be quite confusing when a customer needs to be compliant and is unable to easily determine how many cpu sockets are in the system via webui.

Comment 6 Barnaby Court 2018-10-15 15:40:28 UTC
Moving to the subscription management component as the issue is in the foreman DB per comment 5.


Note You need to log in before you can comment on or make changes to this bug.