Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 2099997

Summary: Adding a new ISCSI storage connection you always get tpgt=1
Product: [oVirt] ovirt-engine Reporter: Avihai <aefrat>
Component: BLL.StorageAssignee: Arik <ahadas>
Status: CLOSED DUPLICATE QA Contact: Avihai <aefrat>
Severity: high Docs Contact:
Priority: unspecified    
Version: 4.5.1CC: bugs
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2022-06-27 14:22:20 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Storage RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Video_Capture_and_engine_vdsm_logs none

Description Avihai 2022-06-22 07:22:12 UTC 
Created attachment 1891811 [details]
Video_Capture_and_engine_vdsm_logs

Created attachment 1891811 [details]
Video_Capture_and_engine_vdsm_logs

Description of problem:
Add a new ISCSI storage connection via UI, you always get tpgt=1.
Remove - re-add an existing storage domain, you always get tpgt=1 (although prior tpgt !=1)

This causes host to add a faulty and duplicate node to ISCSI DB which does not align with actual storage tpgt(which is usually not =1) which results in issues like bug 2097614 when host goes to non-operational state as the engine belives the host does not have connectivity to the engine although it does.

Version-Release number of selected component (if applicable):
rhv-4.5.1-3


How reproducible:
100%

Steps to Reproduce:
1. Have an existing ISCSI storage domain
2. Via webadmin: add a new storage connection

Webadmin -> Storage -> Stand on/choose existing iscsi storage domain -> press the "Connections" bottom -> Press "Add" -> Insert Address,Port(default iscsi is 3260), Target (iqn.XXX:YYY) -> Press checkmark "V"

Actual results:
Checkout the engine DB storage connection[1] and see you get portal(which is tpgt) value of '1'.

[1]
> /usr/share/ovirt-engine/dbscripts/engine-psql.sh -c "select connection,iqn,port,portal from storage_server_connections;

connection | iqn        | port | portal                                               <IP/FQDN>  | iqn.XX:YYY | 3260 | 1


Expected results:
Allow for adding other than '1' value for adding new storage connections.



Additional info:
The new connection was defined using the webadmin
Comment 1 Arik 2022-06-22 07:31:59 UTC 
Avihai, considering that bz 2099640 moved to ovirt-engine and that it shouldn't have been reported on RHV, is there a reason to keep this one or close as duplicated?
Comment 2 Avihai 2022-06-22 07:53:44 UTC 
Hi Arik,
This new bug was requested by @Nir and I agree with it.

From my and Nir's understanding: 
bug 2099640 should fix engine behavior of user adding the wrong tgpt value(via UI/API/Ansible) - we don't want host to become non-operational.

This bz is about fixing storage connection feature which currently any new added storage connection is added with tpgt=1.

I think having a separate bug is the right and more clear way to resolve each issue separately , but I leave final decision up to you ofcs.
Comment 3 Arik 2022-06-22 08:04:50 UTC 
sorry, I still don't understand
we have three bugs:
1. that should prevent the host from becoming non-operational by VDSM not reporting an error on duplicate connections
2. bug on Ansible
3. bug on Webadmin

so you say that 2 and 3 are unrelated to the source that the request comes from but whether tgpt was manually specified or not?
I didn't see any indication that someone explicitly specified incorrect tgpt on those bugs so if that's the case, then they need to be updated - first to reflect the manual specification of tgpt and second, move both to ovirt-engine
Comment 4 Arik 2022-06-22 08:14:56 UTC 
or do you mean this is an RFE to be able to specify tgpt from the UI?
Comment 5 Avihai 2022-06-22 13:01:49 UTC 
(In reply to Arik from comment #3)
> sorry, I still don't understand
> we have three bugs:
> 1. that should prevent the host from becoming non-operational by VDSM not
> reporting an error on duplicate connections
> 2. bug on Ansible
> 3. bug on Webadmin
> 
> so you say that 2 and 3 are unrelated to the source that the request comes
> from but whether tgpt was manually specified or not?
Not sure I follow the question but after the discussion we have I hope it's cleared up.

> I didn't see any indication that someone explicitly specified incorrect tgpt
> on those bugs so if that's the case, then they need to be updated - first to
> reflect the manual specification of tgpt and second, move both to
> ovirt-engine

After our discussion we agreed that user should not be able to add the ISCSI portal/tgpt and engine should do that as it does when adding a new storage domain via the discovery.
This change can be resolved in this bug.

Changing Expected results:
User should see in the engine DB and host ISCSI node DB the real ISCSI portal/tgpt and not automatically portal/tgpt =1.
Comment 6 Arik 2022-06-27 14:22:20 UTC 
Similar flow to that described in bz 2099640 but from the webadmin - if we introduce a fix on the backend side, it should cover both flows

*** This bug has been marked as a duplicate of bug 2099640 ***