(RHEL-17394) fstab-generator: Chase symlinks where possible #150
Conversation
This has a long history; see see 5261ba9 which originally introduced the behavior. Unfortunately that commit doesn't include any rationale, but IIRC the basic issue is that systemd wants to model the real mount state as units, and symlinks make canonicalization much more difficult. At the same time, on a RHEL6 system (upstart), one can make e.g. `/home` a symlink, and things work as well as they always did; but one doesn't have access to the sophistication of mount units (dependencies, introspection, etc.) Supporting symlinks here will hence make it easier for people to do upgrades to RHEL7 and beyond. The `/home` as symlink case also appears prominently for OSTree; see https://ostree.readthedocs.io/en/latest/manual/adapting-existing/ Further work has landed in the nspawn case for this; see e.g. d944dc9 A basic limitation with doing this in the fstab generator (and that I hit while doing some testing) is that we obviously can't chase symlinks into mounts, since the generator runs early before mounts. Or at least - doing so would require multiple passes over the fstab data (as well as looking at existing mount units), and potentially doing multi-phase generation. I'm not sure it's worth doing that without a real world use case. For now, this will fix at least the OSTree + `/home` <https://bugzilla.redhat.com/show_bug.cgi?id=1382873> case mentioned above, and in general anyone who for whatever reason has symlinks in their `/etc/fstab`. (cherry picked from commit 634735b) Resolves: RHEL-17394
In case fstab-generator is called in the initrd, chase_symlinks() returns with a canonical path "/sysroot/sysroot/<mountpoint>", if the "/sysroot" prefix is present in the path. This patch skips the "/sysroot" prefix for the chase_symlinks() call, because "/sysroot" is already the root directory and chase_symlinks() prepends the root directory in the canonical path returned. (cherry picked from commit 98eda38) Related: RHEL-17394
…ntries (#8281) Let's make this LOG_DEBUG, as this didn't used to be an issue, and shouldn't really be still. Replaces: #8132 (cherry picked from commit f1a2c75) Related: RHEL-17394
Based on: 0307ea4 Related: RHEL-17394 rhel-only
Otherwise --template= simply doesn't work: $ ./systemd-escape --template=systemd-mkfs@.service --path /dev/test9 is_valid: 0 is_template: 1 Template name systemd-mkfs@.service is not valid. Related: RHEL-17394 rhel-only
According to bootup(7) and the behavior when /usr is specified in /etc/fstab, the /sysroot/usr mount should be before initrd-fs.target, not before initrd-root-fs.target. Related: RHEL-17394 (cherry picked from commit 104bc12)
|
I added a couple of commits that backport the "fstab-generator" part of TEST-81-GENERATORS from upstream. SInce its behavior differs a lot from the current upstream, the code is littered with comments about what either doesn't work at all or works in a different way in RHEL 7. Also, the test apparently discovered one real bug, so I backported the fix as well. And for completeness, the test fails as expected if 634735b is cherry-picked without the follow-up. |
mrc0mmand
force-pushed
the
rhel-7.9
branch
2 times, most recently
from
c83846d
to
e16e0dc
Compare
Some fstab-generator features are not present on RHEL 7 or they behave differently - in such case there's an inline comment explaining what's different with a reference to an upstream commit that introduced the changed behavior. Also, RHEL 7 systemd doesn't allow overriding (/sysroot)/etc/fstab or /proc/cmdline, so instead of backporting another bunch of potential risky commits, let's temporarily bind-mount a modified copy of necessary files in place of the expected ones. One exception is $SYSTEMD_IN_INITRD, since systemd checks if the mount for / is a tmpfs, which is a pain to mock, but the patch for that is, thankfully, pretty small. Related: RHEL-17394 rhel-only
Commit validationTracker - RHEL-17394 The following commits meet all requirements
Tracker validationSuccess🟢 Tracker RHEL-17394 has set desired product: Pull Request validationSuccess🟢 CI - All checks have passed Auto MergeFailed🔴 Pull Request has unsupported target branch Success🟢 Pull Request is not marked as draft and it's not blocked by |
|
Would this be a good time to revisit this PR? |
|
@msekletar and @dtardon, could you please have a look? Thanks |
|
@mrc0mmand Are you sure the fstab-generator test actually tests anything? AFAICS it requires systemd/systemd@ed4ad48, systemd/systemd@99e3d47, systemd/systemd@0307ea4, and systemd/systemd@9a135c0. |
Yes, I'm very sure, see the code. The |
All right, I missed that. |
github-actions
bot
added
pr/needs-manual-merge
and removed
pr/needs-review
Backport of three fstab-generator commits:
commit 634735b ("fstab-generator: Chase symlinks where possible (#6293)" -- reverted in PR#149)
commit 98eda38 ("call chase_symlinks without the /sysroot prefix (#6411)" -- fixing the issue that required the revert)
commit f1a2c75 ("fstab-generator: downgrade message when we can't canonicalize fstab entries (#8281)" -- adjust log message priority)
Resolves: RHEL-17394