The following proposal comes out of the discussion at this weeks Server SIG
meeting[1]
Fedora Server will have:
* / (root) will be a minimum of 2 GiB and a maximum of 15 GiB
* SWAP will continue to be calculated automatically based on available RAM on
the system
* All unused space will be assigned to a volume group and available to be
assigned to new partitions or extend existing partitions.
* Anaconda will continue to handle the appropriate EFI and /boot settings
We also discussed during the meeting whether we should have a separate /var
partition by default, but the general sense was that we might be better served
by developing a mechanism to allow partitions to be split from existing mount
points, which would be more flexible going forward.
As we did not have quorum in the meeting by the point we got to this proposal,
I'm taking it to the list for discussion and votes.
For the record, the current behavior of the partitioning scheme is for / to be
given up to 50 GiB of space and then any remaining space after that is assigned
to a separate /home partition.
[1]
https://meetbot.fedoraproject.org/fedora-meeting-1/2016-03-15/serversig.201…
Oh wow, it's already that time in the Fedora 31 release cycle. Welcome
to the first Fedora 31 Beta blocker status email!
Action summary
====================
Accepted blockers
-----------------
1. bes — https://bugzilla.redhat.com/show_bug.cgi?id=1728419
ACTION: Server WG to decide if it makes sense to include "a DAP
server" in the optional packages for the web-server group.
2. libdnf — https://bugzilla.redhat.com/show_bug.cgi?id=1727343
ACTION: QA to verify libdnf-0.35.1-2 fixes the bug
Proposed blockers
-----------------
1. anaconda — https://bugzilla.redhat.com/show_bug.cgi?id=1732620
ACTION: anaconda maintainers to build new package which includes upstream fix
2. gnome-shell — https://bugzilla.redhat.com/show_bug.cgi?id=1731908
ACTION: gnome-shell maintainers to verify the issue is resolved after rebuild
3. initial-setup — https://bugzilla.redhat.com/show_bug.cgi?id=1731415
ACTION: initial-setup maintainers to diagnose issue
4. kernel — https://bugzilla.redhat.com/show_bug.cgi?id=1733388
ACTION: QA team to determine if btrfs should be excluded from blocker
criteria and perhaps propose dropping it from the installer
5. systemd — https://bugzilla.redhat.com/show_bug.cgi?id=1715699
ACTION: decide if this qualifies as a blocker
Bug-by-bug detail
=============
Accepted blockers
-----------------
1. bes — https://bugzilla.redhat.com/show_bug.cgi?id=1728419 — NEW
bes not installable: needs rebuild, probably update to 3.20.5
Package was added to comps to replace dap-server package which no
longer exists. It's unclear if the comps group should continue to
carry bes as an optional package. If it can be
dropped, this is no longer a blocker. If it should remain, the bes
package needs to be fixed.
2. libdnf — https://bugzilla.redhat.com/show_bug.cgi?id=1727343 — ON_QA
PackageKit: libdnf::Repo::Impl::detachLibsolvRepo(): packagekitd
killed by SIGSEGV
Appears to be fixed in libdnf-0.35.1-2.
Proposed blockers
-----------------
1. anaconda — https://bugzilla.redhat.com/show_bug.cgi?id=1732620 — MODIFIED
FreeIPA enrolment via kickstart fails since
Fedora-Rawhide-20190722.n.1 (anaconda-31.20-1.fc31), 'realm join' step
not run at all
Fix merged upstream: https://github.com/rhinstaller/anaconda/pull/2052
2. gnome-shell — https://bugzilla.redhat.com/show_bug.cgi?id=1731908 — NEW
After upgrading gnome-shell, mutter and gnome-desktop3 to version
3.33.4-1 disappear login screen
Suspected to be related to a soname bump in libgnome-desktop. If so,
it will be resolved when components are rebuilt against the new
version.
3. initial-setup — https://bugzilla.redhat.com/show_bug.cgi?id=1731415 — NEW
Network spoke doesn't work on ARM initial-setup
Selecting the "network setup" spoke causes installation to hang on
Raspberry Pi 3.
4. kernel — https://bugzilla.redhat.com/show_bug.cgi?id=1733388 — NEW
Circular locking often causes btrfs installs to hang with
kernel-5.3.0-0.rc0.git7.1.fc31 and later
Installations since kernel-5.3.0-0.rc0.git7.1.fc31 often fail during
package installation. The kernel team proposes excluding btrfs from
any blocker criteria.
5. systemd — https://bugzilla.redhat.com/show_bug.cgi?id=1715699 — POST
LiveOS boot, journalctl is missing many early messages
Rejected as a blocker previously, but renominated. The question is:
does missing the first ~25 seconds of log messages violate the "is
working" part of the logging criterion?
--
Ben Cotton
He / Him / His
Fedora Program Manager
Red Hat
TZ=America/Indiana/Indianapolis