Thanks to everyone who participated in the discussion this week.
video recording: https://www.youtube.com/watch?v=Zhx17DyCWFo
meeting notes (thanks Serapheim):
-
What is on the current roadmap in terms of OpenZFS releases?
-
2.0.1 and subsequent 2.0.X releases are coming out soon that better
support the latest Linux kernel versions.
-
2.1 is expected to be released in this quarter
-
Main projects that are inflight that we hope to release in 2.1:
-
DRAID, Forced Pool Export, DirectIO
-
Further quarterly releases are expected to mostly have bug fixes
-
We expect OpenZFS 3.0 within the year~18 months (first half of 2022)
-
RAIDZ Expansion
-
Work is still underway, expect to see progress on that soon
-
Currently Matt is breaking out some ABD related changes from his
branch and a PR with them is expected to be out later this week. These
changes mostly make it possible for the abd_t to be allocated by the
caller. The BSD-specific parts of this work are a bit more
tricky but still
doable.
-
Marking vdevs as non-allocatable
-
Mark Maybee is expected be working on this feature
-
Goal: Making the removal of multiple devices easier
-
Ensure that all devices can be removed (e.g. there is enough space
in the pool for the data after their removal)
-
Don’t spread data of device being removed to devices that will be
removed next, mark them as non-allocatable first as you queue up their
removal.
-
Current design just passivates the vdev’s metaslab group
-
Christian also threw the idea of removing the vdev from the metaslab
class altogether, instead of just passivating the group
-
Allan Jude has some code for the vdev-properties he could try and
upstream soon
-
No-alloc could be a part of this, and save Mark from having to
write an interface to the userland for it
-
Potential Talks with Seagate SCSI/Firmware Folks about priority policies
-
Seagate Firmware Folks want to get a better understanding of how ZFS
prioritizes operations, and coordinate with the community on
their designs
of how operations can be prioritized best.
-
From our community’s perspective, getting a high-level description of
what the driver can do, how we can take advantage of that, and if the
benefit is enough for us to implement any logic, are the main questions.
-
Alexander Motin’s work/discussion of separating priorities for
different requests (ZFS background work, user-triggered requests,
synchronous vs asynchronous, etc..) is also relevant on the matter.
-
For next steps on this:
-
It would be nice to have some information from the firmware folks
before setting a meeting, so we can have a better
understanding of what
this entails and gather up the right folks that can discuss
the topic in
more detail.
-
For setting up the actual meeting, we can coordinate on Slack,
have an initial discussion there, and then send out an email
when everyone
is prepared to talk about this.
On Mon, Jan 4, 2021 at 9:10 AM Matthew Ahrens <mahr...@delphix.com> wrote:
> The next OpenZFS Leadership meeting will be held tomorrow, December 8,
> 1pm-2pm Pacific time. We don't have many topics on the agenda for
> tomorrow's meeting, so let me know if you'd like to add anything.
>
> Everyone is welcome to attend and participate, and we will try to keep the
> meeting on agenda and on time. The meetings will be held online via Zoom,
> and recorded and posted to the website and YouTube after the meeting.
>
> For more information and details on how to attend, as well as notes and
> video from the previous meeting, please see the agenda document:
>
>
> https://docs.google.com/document/d/1w2jv2XVYFmBVvG1EGf-9A5HBVsjAYoLIFZAnWHhV-BM/edit
>
> --matt
>
------------------------------------------
openzfs: openzfs-developer
Permalink:
https://openzfs.topicbox.com/groups/developer/Td6ebc6df2e2a14bd-M837088846c28dd7ed6702ba0
Delivery options: https://openzfs.topicbox.com/groups/developer/subscription
