Re-evaluate the list of release-blocking storage interfaces #196

Open
opened 2026-05-11 09:53:18 +00:00 by psklenar · 3 comments

Fedora Quality team would like to bring up a topic for discussion at your next meeting:

Re-evaluating release-blocking storage interfaces

Background:

Question for Server WG:
Looking at the storage interfaces in our installation test matrix:
https://fedoraproject.org/wiki/Template:Installation_test_matrix#Storage_devices

Which of these should remain release-blocking for Fedora?
Which could potentially be moved to non-blocking or deprecated?

This will help us update release criteria to match actual testing capacity.

Tracking issue: quality/tickets#890

Fedora Quality team would like to bring up a topic for discussion at your next meeting: **Re-evaluating release-blocking storage interfaces** Background: - We decided to review whether we still need to block on so many storage interfaces https://discussion.fedoraproject.org/t/fedora-quality-scope-reduction-announcement-and-summary/160523 - Quality team has lost capacity to test advanced hardware (SAS, HWRAID, iSCSI, FCoE, Multipath) - This affects our ability to validate release criteria Question for Server WG: Looking at the storage interfaces in our installation test matrix: https://fedoraproject.org/wiki/Template:Installation_test_matrix#Storage_devices Which of these should remain release-blocking for Fedora? Which could potentially be moved to non-blocking or deprecated? This will help us update release criteria to match actual testing capacity. Tracking issue: https://forge.fedoraproject.org/quality/tickets/issues/890
Owner

Status as of meeting May 13, 2026

  • keep blocking NVMe SATA, VirtIO, iSCSI
  • remove blocking: PATA SCSI, firmware RAID
  • discuss and decide HW raid here in ticket
  • SAS not explicitly decided, so keeps blocking for now.
*Status as of [meeting May 13, 2026](https://meetbot.fedoraproject.org/meeting_matrix_fedoraproject-org/2026-05-13/fedora-server.2026-05-13-17.00.html)* * keep blocking NVMe SATA, VirtIO, iSCSI * remove blocking: PATA SCSI, firmware RAID * discuss and decide HW raid here in ticket * SAS not explicitly decided, so keeps blocking for now.
Owner

I would like to continue classifying built-in devices as ‘blocking’. This applies particularly to older servers (around 10 years old) that may still be in use (especially in ‘third world’ regions) and are equipped with SAS and hardware RAID.

This is a strategic argument.

I would like to maintain Fedora’s reputation - along with Linux in general - for working perfectly on older devices. This is relevant to the current debate about replacing Windows OS and the accusation of a deliberate obsolescence strategy by many hardware and software companies. From which Linux - and in my opinion including Fedora - stands out favorably, and is viewed and praised in this regard.

Regarding SAS: Obviously we have some access to this, but limited. So we should try hard to use it.

Regarding HW Raid: The situation is much more complex because of the wide range of models. Maybe we can limit the scope to widely used models in professional HW, e.g. Adaptec (that model I have here, secured from our decommissioned devices archive).

And I'm wondering about Red Hat. Back when we had a project with IBM, I worked with the ITSO in Austin that had a huge bunch of test equipment to test software or reproduce bugs. Red Hat must have something alike?

I would like to continue classifying built-in devices as ‘blocking’. This applies particularly to older servers (around 10 years old) that may still be in use (especially in ‘third world’ regions) and are equipped with SAS and hardware RAID. This is a strategic argument. I would like to maintain Fedora’s **reputation** - **along with Linux in general** - for working **perfectly on older devices**. This is relevant to the current debate about **replacing Windows OS** and the accusation of a **deliberate obsolescence strategy** by many hardware and software companies. From which Linux - and in my opinion including Fedora - stands out favorably, and is viewed and praised in this regard. Regarding SAS: Obviously we have some access to this, but limited. So we should try hard to use it. Regarding HW Raid: The situation is much more complex because of the wide range of models. Maybe we can limit the scope to widely used models in professional HW, e.g. Adaptec (that model I have here, secured from our decommissioned devices archive). And I'm wondering about Red Hat. Back when we had a project with IBM, I worked with the ITSO in Austin that had a huge bunch of test equipment to test software or reproduce bugs. Red Hat must have something alike?
Author

Strategic value of supporting legacy hardware (SAS, HW RAID) for Fedora's reputation is very much understood. But practical testing shows:

  • SAS hardware access is acknowledged but very limited
  • HW RAID testing requires specific models

Automation reality has shifted more to cloud. I also have access to a lab with physical hardware for SAS and HW RAID, but with a very long queue and manual step to borrow this hw. As a Fedora QA member - I will always try my best to find such hardware for testing, but making it release-blocking is too much given these constraints.

Strategic value of supporting legacy hardware (SAS, HW RAID) for Fedora's reputation is very much understood. But practical testing shows: - SAS hardware access is acknowledged but very limited - HW RAID testing requires specific models Automation reality has shifted more to cloud. I also have access to a lab with physical hardware for SAS and HW RAID, but with a very long queue and manual step to borrow this hw. As a Fedora QA member - I will always try my best to find such hardware for testing, but making it release-blocking is too much given these constraints.
pboy added this to the Maintenance project 2026-05-21 12:38:37 +00:00
pboy self-assigned this 2026-05-21 12:39:15 +00:00
Sign in to join this conversation.
No milestone
No project
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
server/tickets#196
No description provided.