IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Basically we're choosing the libdnf behavior of omitting `Epoch: 0` which IMO is
the sane thing to do even though in *theory* they're different RPMs but
who would ever ship both without incrementing release/etc?
More information in: https://github.com/projectatomic/rpm-ostree/pull/187Closes: #1198
Approved by: cgwalters
This fixes a large swath of compatibility issues, for the same reasons as
overlayfs makes a lot of things Just Work. The ugly part of course is
doing hidden copyups inside the filesystem.
We've gone quite a long time with the "pure rofiles" mode, and have made changes
to various bits of userspace to be compatible with it. But what finally made me
give up on that is glibc's locale-archive; there's a patch for it that
is stalled, but even if it was applied we would still need to work with
older glibc.
This issue comes to the fore in unified core 🌐 mode, as without this
we won't get a correct locale archive.
Closes: #1171
Approved by: jlebon
This is the first step towards unifying how we introspect packages from
a specific commit. We currently do this in three ways: libdnf, librpm,
and now `rpmostree.rpmdb.pkglist`. I'd like to get to a point where we
only have `rpmostree.rpmdb.pkglist` and libdnf, the latter only when
more complex queries are required.
This patch teaches the `db diff` command to make use of the new db diff
API so that it can work even on metadata-only commits. This is relevant
for use cases mentioned in #558.
I didn't get rid of the `rpmhdrs_diff` functions right now because of
the `--changelogs` option: libdnf currently does not expose this, so we
fall back to the previous API in that case. OTOH, I wonder how much it's
actually used in the wild; maybe we could just nix it?
Closes: #1162
Approved by: cgwalters
The new `rpmdb.pkglist` metadata is a cheap way of retrieving the set of
packages in a commit. I'd like to make use of it as much as possible vs.
checking out the rpmdb and setting up a DnfSack.
Of course, in the case of layered commits, it doesn't matter *as* much,
because a layered commit being present in the repo should mean that a
deployment is currently using it, and we should learn to reuse the rpmdb
checkout of that deployment. Though keeping it consistent across both
server and client commits makes implementing `OstreeDeployment`-agnostic
things like `db diff` more efficient too. I also plan to use this in the
upcoming auto-update code.
Closes: #1158
Approved by: cgwalters
This rolls up several libglnx changes: https://github.com/GNOME/libglnx/pull/101
Now of course things are trickier here because we have an internal
abstraction over directly emitting to a console versus sending the
result over DBus. Further complicating things is that some things
call into libdnf and thus *require* use of `DnfState` which does
not give us the "n items" information, versus other parts which
we implement and can do what we want.
Even *further* complicating things is that we have to take care around non-CLI
callers like Cockpit; so I didn't try to pass the "n items" over DBus, rather
just reimplemented the "insert into text" that libglnx is doing.
Anyways overall this looks better IMO for all cases.
Update submodule: libglnx
Closes: #1143
Approved by: jlebon
This function is now basically reimplementing all of the intelligence
we have in libostree today. It wouldn't have worked before we
fixed https://github.com/ostreedev/ostree/issues/1165 but now
that's done, conceptually we just need to call
`ostree_repo_commit_modifier_set_sepolicy()`. However, I had
to drop the "number of files changed" since currently the libostree
API doesn't support that.
Also, in order to detect the case that content changes at all (so
we still have some useful information in the journal), implement
a "content hash" check.
See also https://github.com/projectatomic/rpm-ostree/pull/1123
(And note we *don't* use `DEVINO_CANONICAL` here)
Closes: #1138
Approved by: jlebon
Right now the fact that one can only cancel via `Ctrl-C` of an existing client
process is rather frustrating if for example one's ssh connection to a machine
drops. Now, upon reconnecting, one can easily `rpm-ostree cancel` a hung update
or whatever rather than doing the more forcible `systemctl stop rpm-ostreed`
(which is safe of course, unless livefs is involved).
Closes: #1019
Approved by: jlebon
Note this PR requires [bubblewrap 0.2.0](https://github.com/projectatomic/bubblewrap/releases/tag/v0.2.0).
Change our bwrap invocations drop truly dangerous capabilities like
`cap_sys_admin` and `cap_sys_module` just like Docker does today. Because of the
popularity of Docker, we can be pretty sure that most RPM scripts should have
adapted to this (although a problematic area here is that traditional librpm
doesn't actually error out if scripts fail).
There are two reasons to do this:
- We want "offline" updates by default; updates shouldn't affect the
running system. If we prepare the new root in the background, a
%post shouldn't restart a service for example. We already "handle"
this by making `systemctl` a symlink to `/bin/true`, but this approach
also shuts off `%post`s that do e.g. `insmod`.
- Protection against accidental system damage
Closes: #1099
Approved by: jlebon
We originally needed the pkgcache to be a separate repo due to ostree's
overzealous pruning policies. The idea was to maintain multiple commits
in each pkg branch for different SELinux policies. In practice, there's
not much use in maintaining old copies and it's just easier to always
relabel on the fly. So then, the need for a separate repo completely
melts away.
This helps simplify the mental model a bit and allows us to avoid subtle
issues like #1047. Note however that the core is still capable of
handling split repos for the `--ex-unified-core` compose use case. Once
that and the jigdo work are a bit more settled, we can have a clearer
picture of how to simplify the core further.
The tricky bit is migrating the cache. When deploying, we check if a
pkgcache repo exists and migrate its refs if so. We then leave behind a
symlink to the system repo to remain compatible with older rpm-ostrees.
Closes: #1055
Approved by: cgwalters
It took me way, way, way too long to debug that my dev container somehow missing
`ostree-grub2` caused the script to fail early, but we'd still continue trying
to sync, leading to us not actually changing the installed `rpm-ostree` binary,
leading to test suite failures.
Also add the chmod hammer at the top so we can delete even if we fail after
doing an rsync.
Closes: #1115
Approved by: jlebon
We have a *lot* of experimental functionality. I think the
`override` bits are fleshed out enough now that we can lift
the `ex` designation. For example, jlebon fixed SELinux
labeling in the presence of override-replace.
Closes: #1089
Approved by: jlebon
Refined a bit for the previous written kargs test.( mainly for
checking kargs after rollback )
Added tests for import-proc-cmdline and deploy-index option
Closes: #1013
Approved by: cgwalters
Refactor the similar code portion from delete option into
a new function, and reuse it in the replace logic.
Add the replace functionality to allow kernel arguments
replacement. Now support two different types of replacement:
1: replace by key=value when there is only one key existing
e.g we have 'hi=hello' as our one of the arguments (hi only has
one value associated with it) , and 'rpm-ostree ex kargs --replace hi=new'
will replace the value and change hi=hello into hi=new.
2: replace by key=value=new_value for all other replacements
( this will work for swapping single value pair too !)
Some tests for rpm-ostree ex kargs --replace are added
for future regression.
Closes: #1013
Approved by: cgwalters
Tests are introduced in this commit to test the basic
functionality for rpm-ostree ex kargs --append,
and rpm-ostree ex kargs --delete.
Those tests are added for future regression.
Closes: #1013
Approved by: cgwalters
Related to: https://github.com/projectatomic/rpm-ostree/issues/774
We aren't yet trying to render this in any way on the client side, but let's
start capturing the timestamp data now so that we can make use of it later
in e.g. `rpm-ostree status -v`.
Closes: #1072
Approved by: jlebon
Depends: https://github.com/ostreedev/ostree/pull/1296
As I mention in the commit there, I see two uses for this in rpm-ostree; first
in our test suite, and second for OCI-built image imports.
I also took a step further here and inject an `original-origin` metadata
key, though we aren't actually using that yet. The problem I'm trying
to solve there is that repeated `make vmoverlay` starts chaining things up,
but that gets very confusing. I think we should always have `vmoverlay` unwind
back to the base ref. (Or at least do that by default)
Closes: #1069
Approved by: jlebon
In another PR I did the manual bridging of commit metadata to deployment
property, but that's annoying. Let's just bridge all commit metadata.
Closes: #1069
Approved by: jlebon
To complement the new `--cache-only` option, add a `--download-only`
option. This does exactly what it says: we download the ostree, download
and import packages, but don't actually commit & deploy. This can be
used to effectively prime a follow-up `--cache-only` operation that can
be done during a more convenient/safer maintenance window.
I debated naming the two options `--pull-only` and `--deploy-only` like
the ostree equivalents. Though "pull" felt like the wrong word given
that it's associated more with ostree pulling but rpm-ostree also
downloads & imports RPMs. As for `--deploy-only` vs `--cache-only`, it
seems like `--cache-only` is a more accurate description of the
functionality (i.e. rather than describing an action, it describes a
mode). I also considered `--no-download` to make the synergy with
`--download-only` more obvious. Maybe that's better? Naming is hard...
Closes: #713Closes: #1049
Approved by: cgwalters
Now that we have a strong notion of `cache-only` mode, make use of it
when performing an `uninstall` or `ex override remove/reset`.
Closes: #944Closes: #1049
Approved by: cgwalters
As Colin mentioned in #1035, the new `--cache-only` implemented only the
rpmmd half of the story. Here we complete that story by also ensuring
that when in cache-only mode, we don't download new ostree data nor new
packages. We try to complete the requested operation with what we have.
To do this, we add support for the same `SYNTHETIC` pull that was added
in ostree[1] so that we don't actually pull, but still perform timestamp
checking.
On the pkgcache side, we disable all remote repos and instead insert all
our cached RPMs into the `DnfSack`. Care is taken to still perform
SHA256 verification for local pkg installs/replacements.
[1] https://github.com/ostreedev/ostree/pull/642Closes: #687Closes: #1049
Approved by: cgwalters
The new idle exit behaviour is nice, but it makes debugging it harder
because you have to be fast enough to attach or place your breakpoints
and trigger it before it auto-exits. Add a compile-time flag that
developers can easily turn on to disable the auto-exit behaviour.
Closes: #1052
Approved by: cgwalters
Just taking what I learned from #1035 and applying it here. What's nice
about this is that there's no cleanup needed. Once the process is killed
(or worst case, we reboot the VM), there's no traces left at all.
Also added a few extra "ok" outputs.
Closes: #1043
Approved by: cgwalters
I've been lazy about actually using using rsync instead of scp when
copying new RPMs over to the VM. We do this here. Also make
`vm_send_test_repo` take a mode argument that allows callers to
completely skip the sending of the repo file itself. This will be needed
for the `makecache` test, in which we *don't* want the repo to be local.
It looks cleaner anyway for the gpgcheck use case as well.
Closes: #1035
Approved by: cgwalters
Prep for implementing `rpm-ostree cancel`, but this works with the way we handle
`Ctrl-C` interactively on a client as well. Being able to cancel a script
execution is quite nice; some of them are expensive, and having one loop forever
has been known to happen.
Closes: #1025
Approved by: jlebon
There's a lot of paths in the core related to SELinux policy changes and
relabeling packages. We currently have no test coverage for them. We add
support in the test libraries here to build such packages.
We also add a test that checks both that we correctly relabel RPMs when
the policy changes and that we handle layered packages that install
SELinux packages properly.
Closes: #999
Approved by: cgwalters
Rather than just letting the scriptlets inherit the daemon's
stdout/stderr, redirect their outputs so that we can set a customized
identifier to make it easier to distinguish from the daemon output.
Also print out the `journalctl` command needed so that users can
investigate the output themselves.
Closes: #998
Approved by: cgwalters
I suspect a common pattern with local replacement overrides is to
simultaneously replace a group of packages that depend on each other in
one shot, as is the case with docker, docker-common, and
docker-rhel-push-plugin currently in Fedora Atomic Host. In such cases,
we can print a cleaner diff in the status to make it easier to grok.
Before:
ReplacedBasePackages: strace 4.18-1.fc26 -> 4.19-1.fc26, docker-common 2:1.13.1-22.gitb5e3294.fc26 -> 2:1.13.1-21.git27e468e.fc26, docker 2:1.13.1-22.gitb5e3294.fc26 -> 2:1.13.1-21.git27e468e.fc26, docker-rhel-push-plugin 2:1.13.1-22.gitb5e3294.fc26 -> 2:1.13.1-21.git27e468e.fc26
After:
ReplacedBasePackages: docker-common docker docker-rhel-push-plugin 2:1.13.1-22.gitb5e3294.fc26 -> 2:1.13.1-21.git27e468e.fc26, strace 4.18-1.fc26 -> 4.19-1.fc26
Closes: #1004
Approved by: cgwalters
We were directly bind mounting the checked out `/usr/etc` onto `/etc`
which was exposing us to corruption from scriptlets. Since we already
have an rofiles-fuse mount for `/usr`, let's just re-use its `etc/`
subdir and bind mount that instead.
Closes: #1003
Approved by: cgwalters
We don't need those in the tree, so let's nuke them. This also fixes
subtle compatibility issues between hardlinks and lock files (see #999).
Closes: #1002
Approved by: cgwalters
This PR uses https://github.com/ostreedev/ostree/pull/1156
to ensure that when installing packages containing files with
exact same content, the files will be merged.
When installing packages containing conflicting files,
the error will still be detected at the ostree side.
The checkout overwrite option at rpm-ostree side is also
modified to accomodate the changes made in ostree side.
A test is added for regression
Commands like `upgrade` and `deploy` need to know if a new deployment
was actually laid down so that it may print a pkg diff if so. This is
implemented by listening for changes to the DefaultDeployment D-Bus
property. D-Bus emits a signal when the deployment variant changes
value.
However, in #595, with the introduction of `pending-*` related keys, the
deployment variant no longer represents data solely tied to that
specific deployment. In this case, because `deploy` operations currently
set the ref to the resolved checksum, it can happen that deploying the
same base commit when the current refspec *isn't* pointing to that base
commit will result in the `pending-*` keys dropping out and a default
deployment change notification going out.
In this patch, we strengthen how we determine whether a new deployment
was laid down by actually looking at the deployment id, rather than just
assuming that a change to the property implies a new deployment.
Closes: #981Closes: #984
Approved by: cgwalters
Also something I noticed while working on #981. When sitting on a livefs
commit, once a user does `rpm-ostree cleanup --pending --rollback`, it's
impossible to redeploy the same booted commit. Let's allow users to do
this.
Closes: #984
Approved by: cgwalters
Today in Fedora the `glibc-all-langpacks.posttrans` is implemented
in lua, for no good reason. See:
https://bugzilla.redhat.com/show_bug.cgi?id=1367585
Since that's stalled out, let's add support for overrides. This
is obviously a much bigger step with more long term maintenance
implications over our current "ignore scripts" list. But we can't
block either.
This is needed for unified core work:
https://github.com/projectatomic/rpm-ostree/issues/729
(We also override `fedora-release-atomichost` but I'll likely
submit a patch for that upstream)
Closes: #980
Approved by: jlebon
I was linking to this code from elsewhere and noticed that
for our hardlink breaks we were not using fd-relative even
though we can. Down the line if we fork librpm into a separate
process and do e.g. `--dbpath=.` it'll do it too.
(Side note, I verified that commenting out the hardlink breaking
here was caught by the `ostree fsck` I added to the test suite)
Closes: #979
Approved by: jlebon
This fixes `rpm-ostree reload` as root, and supports configuring
it to be enabled for other users as well. This was overlooked
in the polkit work originally.
Closes: https://github.com/projectatomic/rpm-ostree/issues/976Closes: #977
Approved by: jlebon
Let's make SELinux and atomic-host-tests happy with trees we concoct
ourselves by using the new --selinux-policy. (Specifically, we want our
sync'ed binaries to have install_exec_t).
Closes: #953
Approved by: cgwalters
I had meant for this to be in the other PR#968. I originally did both
`make` and `make install` there, but now it only does `make install`, so
let's just rename it to make that more obvious.
Closes: #953
Approved by: cgwalters
At one point `rpm-ostree install libvirt` dragged in libguestfs which in turn
brought in `syslinux-extlinux-nonlinux` which has files in `/boot/extlinux`,
which we rejected. (That dependency chain appears to have been fixed currently)
For the general case, this is just a partial fix in that we haven't nailed down
the semantics of how updates for `/boot` work. But in this particular case,
we'll just break libguestfs' `extlinux` verb, which I'm OK with.
Another case is `fwupdate-efi` - we require manual intervention to copy the
data into `/boot` after installing the package.
This is also preparation for [unified core](https://github.com/projectatomic/rpm-ostree/issues/729)
in that we now ensure imported kernels don't end up in `/boot` unless
explicitly configured.
Closes: https://github.com/projectatomic/rpm-ostree/issues/853Closes: #969
Approved by: jlebon
Let's make using a custom install tree easier and document the process.
We split out the insttree step into `build.sh` so that we no longer have
to `flock(1)` around it, and also share between `overlay.sh` and
`sync.sh`.
Closes: #968
Approved by: cgwalters
While working on unified core and the Fedora Atomic Host content set, I hit a
dependency between `docker.posttrans` which tries to read `/etc/os-release`, and
`fedora-release-atomichost.post` which creates that symlink.
It seems best practice to me to run `%post`s strictly before
`%posttrans`; we're not likely to do parallelization anytime
soon anyways.
While here I cleaned things up by having an enum for the script kind,
rather than multiple functions, otherwise we would have had another
wrapper in core.c.
Closes: #963
Approved by: jlebon
Since we have a copy of this libostree code, pick up the new
changes from <https://github.com/ostreedev/ostree/pull/1055>.
Note the added test doesn't really test our logic since
we're only doing local pulls, but at least we have something.
Closes: #932
Approved by: jlebon
This closes a longstanding bug - since package layering first
landed, we only checked for newer RPMs if the base tree changed.
In some scenarios like RHELAH, this doesn't matter much by default
since they move at the same cadence. Except if you use EPEL for example.
In Fedora, today the FAH releases are async of the rpm-md repos, and
there's also COPR which can update more than once a day even.
We should check for both update sources. Luckily we'd already introduced logic
for this in the treecompose case (checksumming the depsolved package sack). We
just need to start using it for client side assembly too.
Closes: https://github.com/projectatomic/rpm-ostree/issues/391Closes: #911
Approved by: jlebon
I'm working on getting the vmcheck suite working as part of Fedora's new
CI pipeline. In that context, we want to test the deliverable as it is,
i.e. with SKIP_VMOVERLAY=1. For compatibility with the testsuite, we
ensure that the machine is on the vmcheck branch before starting the
tests.
Eventually, we should try to make the vmcheck suite runnable outside of
a configured build directory to make it easier to re-use in such
contexts.
Closes: #917
Approved by: cgwalters
The new API to find pending and rollback deployments do so relative to
the booted deployment. This caused an interesting behaviour: the first
time a user uses "rpm-ostree rollback", it would (as expected) move the
previous deployment first. but the second call to "rpm-ostree rollback"
would fail since there were now no more rollback deployments.
We fine tune the logic here to allow this, as well as the more general
case of putting the booted deployment back on top.
This fixes a subtle regression from b7cf58e
(https://github.com/projectatomic/rpm-ostree/pull/767).
Closes: https://github.com/projectatomic/rpm-ostree/issues/906Closes: #907
Approved by: cgwalters
Fix regression from https://github.com/projectatomic/rpm-ostree/pull/870
which caused `vmsync` and `vmoverlay` to no longer actually overlay
ostree bits.
We go back to using `--files-from`, but just make sure to filter out the
stuff that we don't need (and which previously caused issues).
Closes: #907
Approved by: cgwalters
When commit metadata contains ostree.endoflife attribute,
its information will be added to the deployment Variant,
which will later be shown as a red & bold message when
'rpm-ostree status' command is called.
A test is added for future regression
Closes: #889
Approved by: cgwalters
File triggers are a post-RHEL7 thing; more information at
http://rpm.org/user_doc/file_triggers.html
There are two notable users I've been testing this with;
`glib2` and `vagrant`. The `vagrant` one is more immediately urgent,
since it makes `vagrant-libvirt` work, which I currently rely on
for my workstation dev.
I've tested things successfully with `vagrant`, and I did verify that we run the
`glib2` ones when doing `rpm-ostree ex container`.
Long term, more transaction file triggers are likely to live in
"base" packages like `glib2`. We don't implement those yet, but
extending this to do that shouldn't be too hard.
There was *significant* what I'd call reverse engineering of the
implementation in librpm. The file triggers code there is spread out
and abstracted in a few different places in the code. I found
trying to understand what header values were involved to be quite
tricky.
There are some corner cases like multiple patterns that I *think*
this does correctly, but could use more validation. The main
question I had was - is it required that the patterns for e.g.
`%transfiletriggerin` and `%transfiletriggerun` be identical?
Closes: https://github.com/projectatomic/rpm-ostree/issues/648Closes: #869
Approved by: jlebon
Follow-up tweak to #894. Make the client smarter so we only register
when we know we can. We could be more sophisticated here and e.g.
introduce the concept of "read-only" clients in the daemon to only allow
access to non-mutating methods, though let's delay that discussion at
least until the daemon learns to auto-exit.
Closes: #898Closes: #900
Approved by: cgwalters
This is a followup to https://github.com/projectatomic/rpm-ostree/pull/888
but more comprehensive; in the layering case, the sanitycheck runs
after all the `%posttrans` scripts, so we'll get a consistent error message
for the `rm -rf /` test.
We also do the sanitycheck for the "pure ostree" case, as well as cases
where we didn't actually layer packages (including `ex override remove` as
well as simply regenerating an initrd).
There's obviously a lot more we could do in a sanitycheck; as I say in the
comment it's tempting to consider trying to boot systemd (in a fully volatile
config), but for now let's do this. In the end of course the admin has rollback
too.
Closes: #892
Approved by: jlebon
Now that we require clients to have an active session to RegisterClient,
we can't use runuser to check for non-root functionality. Add a new
vm_cmd_as() function to allow connecting as a different user. While
we're there, do some minor cleanups to consistently use `local` when
possible.
Closes: #894
Approved by: cgwalters
This test would actually fail even if the bin user were allowed to
install a package because there are no enabled repos to install. Fix it
so that we know we have foo there and explicitly check that the error
message is what we expect.
Closes: #894
Approved by: cgwalters
I was thinking today about our script handling, and I realized
an excellent way to showcase the advancement rpm-ostree makes
over traditional package managers is the fact that we survive a
`%post` script that does `rm -rf /`!
See e.g. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=858521
It's been this way ever since we first implemented package layering;
the fact that we construct a new root and use bubblewrap to sandbox
makes us very resilient to this type of thing.
But, let's add a test case for this to be sure we preserve this behavior; for
example, if in the future we for some reason we decide to leak some host state
into the scripts.
Closes: #888
Approved by: jlebon
This is required for glibc-all-langpacks at least:
https://bugzilla.redhat.com/show_bug.cgi?id=1367585
Otherwise, its usage is...extraordinarily rare. In fact looking at a snapshot of
`rpm-specs-20170518.tar.xz` from Fedora, the only other use is in
`postfix.spec`, and it appears bogus (the value is already expanded at build
time).
But the glibc case is special, as the value of `install_langs` is indeed
potentially dynamic per system.
Closes: #873
Approved by: jlebon
The `install` command in CentOS 7 is too old to understand that
`-Dt foo/bar` means creating both `foo` and `bar`, which is useful so
that we avoid an explicit `mkdir` before. But we can't do that here.
Closes: #871
Approved by: cgwalters
Add experimental support for replacing packages from the base layer with
local RPMs. This is useful for example, to cherry pick a fixed package,
or to roll back to a previous package version. Like with pkg removals,
only files in /usr are actually replaced.
This patch also contains a few usability improvements as well, e.g.
showing the full NEVRA of removed packages rather than just their names,
and support for resetting overrides using either the pkgname or NEVRA.
Closes: #852
Approved by: cgwalters
Make sure that we wipe out any leftover configuration files from a
previous run before layering the test pkg, or we'll get false positive.
Also make sure to correctly clean up the VM in the case the livefs test
errors out.
Closes: #859
Approved by: cgwalters
Seen in the wild with `vagrant`'s use of `%post -p /usr/bin/ruby`. This was a
very easy fix, and actually makes the code a little bit nicer, as we no longer
need to explicitly make the script executable, since we now pass it as
`argv[1]`, the same way librpm does. That in turn would make it possible to fix
the TODO and use `bwrap --file`, but that can come later.
Closes: https://github.com/projectatomic/rpm-ostree/issues/856Closes: #858
Approved by: jlebon
Nuke all the previous goop that was used to create RPMs at `make check`
time and transition all the tests to use the new `build_rpm` function.
It definitely feels cleaner to use. It's also really nice to have the
spec live in the same file as the test that uses it.
Closes: #854
Approved by: cgwalters
The property of removal overrides dropping out if the package was
removed from the base layer felt a bit too magical and hacky. We really
should remember that wish and re-apply it if the pkg comes back. This is
similar to package layering: requests can become inactive (seems like a
better word than "dormant") if the package is already part of the base
layer, but they don't really go away.
This patch reworks the logic so that removal overrides work the same
way. In the status output, we now have both "RemovedBasePackages" and
"InactiveBaseRemovals" (which is only printed in verbose mode),
similarly to how we have "LayeredPackages" and "InactiveRequests". And
similarly, we also print out in the upgrader during a transaction all
the inactive base removals.
Another cool thing is that we now allow any pattern to be specified at
the CLI. E.g. `ex override remove /usr/bin/strace` will resolve to
strace.
Closes: #836
Approved by: cgwalters
In the error path when trying to remove a base package, we would try to
print a DnfPackage as char*, which of course didn't result in any
coherent output.
Closes: #833
Approved by: cgwalters
To catch things like https://github.com/projectatomic/rpm-ostree/issues/826
during testing.
(Really IMO this should be the system-wide default except GNOME and
lots of apps would start dumping core all over the place... 😢)
Closes: #828
Approved by: jlebon
Copy the bits from `overlay.sh` that takes the ostree files from the
build container so that we also get updated ostree bits when we
`vmsync`.
Also make sure to `daemon-reload` before restarting the daemon.
Closes: #803
Approved by: cgwalters
This was a kind of last-minute bug introduced when I tweaked the
checkout to use `.` to avoid a `mkdir()` for files. But there were
multiple bugs with that; for files that are in subdirectories of `/etc`
we obviously need to get the right subdir and not use `/etc`.
Second, we need to handle the case where we're adding new subdirectories.
This change fixes `rpm-ostree install firewalld + rpm-ostree ex livefs`.
Closes: #783
Approved by: jlebon
This is a valid case when layering local RPMs. Otherwise (e.g. if
specific non-local packages are requested), give the nicer error rather
than letting libdnf throw something less useful.
Closes: #780Closes: #781
Approved by: cgwalters
There are a few different use cases here. First, for layering new packages,
there's no good reason for us to force a reboot. Second, we want some support
for cherry-picking security updates and allowing admins to restart services. Finally,
at some point we should offer support for entirely replacing the running tree
if that's what the user wants.
Until now we've been very conservative, but there's a spectrum here. In
particular, this patch changes things so we push a rollback before we start
doing anything live. I think in practice, many use cases would be totally fine
with doing most changes live, and falling back to the rollback if something went
wrong.
This initial code drop *only* supports live layering of new packages. However,
a lot of the base infrastructure is laid for future work.
For now, this will be classified as an experimental feature, hence `ex livefs`.
Part of: https://github.com/projectatomic/rpm-ostree/issues/639Closes: #652
Approved by: jlebon
Now that rpm-ostree operations are cumulative, we need to make sure we
also clean up pending deployments. Also clean up the rollback for
consistency.
We also strengthen our ref handling. Always nuke the vmcheck ref (which
might not even exist anymore if a test rebased without --skip-purge) and
recreate it from the original ref.
Closes: #769
Approved by: cgwalters
Conceptually: we're going to move rpm-ostree and ostree at the same
cadence most of the time; for both releases *and* for git master.
The problem so far has been the latter part. Reusing FAHC
for the build gets us half of the problem.
The other trick I realized we can do - just pull ostree out from the build
container. This avoids fetching it from the internet, and makes my workflow for
hacking on both nicer - I just `sudo make install` in my build container for
ostree.
It's tempting to make the whole thing symmetric and require `sudo make install`
for rpm-ostree and not do the insttree thing but...perhaps after.
Closes: #758
Approved by: jlebon
This is preparatory for future changes which will make use the newer sysroot
writing API. But in this change, we can drop all of our version ifdef'd code.
Closes: #740
Approved by: jlebon
I want to be able to debug a test by sshing in with the same setup;
right now the suite resets and reboots by default.
This also required fixing how we deal with `/etc/yum.repos.d`, since at the
start of the test it could be either pristine or not.
Closes: #578
Approved by: jlebon
I need to build ostree from git too. So now my workflow is:
```
export insttree=/srv/walters/tmp/rootfs
cd ostree
make && make install DESTDIR=${insttree}
cd rpm-ostree
make && make install DESTDIR=${insttree}
env VMCHECK_INSTTREE=${insttree} make vmoverlay
```
Closes: #705
Approved by: jlebon
I originally thought there was a bug here, but missed we were skipping
symlinks earlier above. See [previous discussion][1].
Now, I'm not aware right now of something this patch actively fixes, but I think
it makes sense to be consistent in our handling of things here with respect to
symlinks.
1: 29dd1bd801..8158dcfb47 (r95017893)Closes: #689
Approved by: jlebon
Allow the `make vmcheck` target to take a HOSTS var, which is simply a
space-separated list of hosts on which we can run testsuites. Add a
multitest.py script that takes care of monitoring and scheduling the
tests onto the nodes.
The script itself is "dumb": we don't know how long each test can take,
so we can't do any smart/heuristic scheduling that could save more time.
Closes: #675
Approved by: cgwalters
1. Don't require an ssh-config
In the case of redhat-ci, the VMs are already fully configured for the
system (injected in the hosts file, host key accepted, etc...). So
there's no need to have an ssh-config there. In general, it should be
acceptable to run the vmcheck suite against a resolvable host without
having to create an ssh-config for it.
2. Make the host name configurable
Rather than hardcoding "vmcheck" as the hostname, allow overridding it
by specifying a VM env var directly. We also prepare the various scripts
to make use of the $VM variable whenever host-specific dirs/files are
created so that parallel runs won't step on each other.
Closes: #675
Approved by: cgwalters
Pull #646 introduced a subtle regression: we went from always including
a "packages" entry to only including it if there are packages present.
Albeit it's easy to guard against, though to be nice, let's make it
easier for consumers by always including it.
Reported-by: Micah Abbott <miabbott@redhat.com>
Closes: #670
Approved by: cgwalters
While reading a recent conversation about GPG checking at treecompose
time, I had a sudden thought - were we actually doing verification
client side? Turned out, we aren't. That happens as part of
`dnf_transaction_commit()` which we don't use.
That function verifies every package at one go, but for us I think it's better
to do it before "importing". We shouldn't have untrusted bits that we've
unpacked (they might have suid binaries, for one thing).
This is an embarassing problem, but it's worth emphasizing that everyone should
be retrieving repodata at a minimum over TLS, which sets a baseline. On RHEL, we
already do pinned TLS, and there are discussions about extending that elsewhere.
See: https://bugzilla.redhat.com/show_bug.cgi?id=1422157Closes: #656
Approved by: jlebon
We don't want to expose the host's `/tmp` since that means scripts could
potentially find things like the X11 socket or whatever.
To debug things better, add a quick bash script to run bwrap like the C code
does. Perhaps down the line we can add `rpm-ostree internals run-bwrap` or so.
Closes: #647
Approved by: jlebon
Clean up the output of vmcheck a bit by printing out ssh debug details
and rpm-ostree status output to the log file. This will help make it
easier to quickly see which tests failed (as well as helping satisfy my
OCD for a nice output).
Closes: #645
Approved by: cgwalters
See https://github.com/projectatomic/rpm-ostree/issues/233 - for RPMs which
place files in e.g. `/opt`, we have different behavior in the treecompose case
(silently drop it) versus package layering (does the wrong thing).
Since the unpacker right now is only used in the layering case, this just
ensures we'll get a consistent error there.
Closes: #624
Approved by: jlebon
I'm not sure why we weren't doing this before, but we need to also
support files in /var and /run that are owned by root.
Related: RHBZ#1421781
Closes: #622
Approved by: cgwalters
We sometimes talk about using `ostree admin undeploy`, but that
doesn't know about the pkgcache, and hence space there leaks
until the next rpm-ostree operation.
Just for this, we need to expose a cleanup command (and API). But
we also need to support cleaning:
- repomd
- downloads (repo/tmp)
So let's start implementing that.
Closes: #614
Approved by: jlebon
Until now, we always used the booted deployment, and would
garbage collect the "pending" deployment. This is the
way OSTree was designed, but I think for rpm-ostree given
how mutable we are on the client side, there's a much stronger
argument for being more stateful too.
This is a relatively simple code change to split the "merge deployment"
concept into two. There's now the "config merge deployment" and the
"origin merge deployment".
Basically, `rpm-ostree install foo; rpm-ostree install bar` will
now install both `foo` and `bar`. But we will still use the booted
deployment for `/etc`.
Down the line, I think I'd like to drive into OSTree the concept of
a "staged" deployment, that has the hardlink checkout done, but doesn't
have the config merge.
But we don't need to change the OSTree core for this yet; we can
do it here in rpm-ostree, and this relatively simple code change
fixes many issues at once.
For example, `rpm-ostree upgrade && rpm-ostree install foo` now
does what you expect as well.
Obviously, we want to enable doing multiple things in *one* transaction,
and we're not far away, but I think this is also the right thing to do now.
I'm relatively confident it won't break anyone's workflow, as what
we did before wasn't generally that useful. However, people will
need to learn to `ostree admin undeploy 0` if they *don't* want
this behavior. (We need to have `rpm-ostree cleanup`).
Closes: https://github.com/projectatomic/rpm-ostree/issues/406Closes: #611
Approved by: jlebon
I'm watching https://github.com/rpm-software-management/libdnf/pull/199 and I
really don't like it. We already have a place to put out-of-rpmdb metadata,
which is in the ostree commit for imported packages. No need to involve a
relational database for this (and further, one that would need to learn about
multiple ostrees).
We're not yet *using* this information in the UI, but we could; imagine
changing the `status` `Packages:` to show packages-per-repo or so. We
could also expose an `rpm-ostree pkg-info foo`.
But for now, let's just start recording this.
Closes: #610
Approved by: jlebon
One thing that's very confusing about OSTree is there are two layers -
deployments and the refs/commits. If one does an `rpm-ostree upgrade`, but then
e.g. `ostree admin undeploy 0`, you still have the new revision in the repo.
We don't do a good job of displaying this state, or helping people clean
it up.
Down the line, I also want to better support something like `rpm-ostree pull` to
cache updates explicitly *without* deploying.
This commit just adds a bit of information to the status display. We might want
to have better formatting, but I think this an OK start.
Closes: #595
Approved by: jlebon
The actual problem I am trying to fix with this is fallout from the
introduction of `/usr/libexec/rpm-ostreed`, which required a SELinux
policy change. Specifically for CentOS, the base policy is rev'd
slowly.
My hope was that by merging the daemon code back into `/usr/bin/rpm-ostree`
which is labeled `install_exec_t`, starting via systemd would do
the right thing. It turns out that doesn't happen.
Now later, I'm picking this patch back up because I want to do multprocessing in
the daemon (and in the core), and it makes sense to share code between them,
because multiprocessing will need to go through a re-exec path.
Another benefit is we avoid duplicated text (libglnx, internal helpers) between
the two binaries.
Closes: #292
Approved by: jlebon
When we checked out the base tree for package layering, we would create
the directory in which ostree did the checkout. This meant however that
ostree wouldn't apply xattrs on the root directory itself. This would
cause the directory to be mislabeled (as system_conf_t instead of
root_t), which in turn cause SELinux violations on reboot when systemd
tried to make the root mount shared.
This patch fixes this by first settling on a permanent directory in
which to do checkouts -- really, we'll never have multiple package
layering operations going on at the same time. Once we know that we have
a reserved path, we can safely let ostree create it for us with the
proper xattrs.
Resolves: RHBZ#1318547
Closes: #605
Approved by: cgwalters
Currently we push for a model where the initramfs is
generated (in non-hostonly mode), and merely replicated.
However, to support a few unfortunate corner cases like dm-multipath which wants
to inject a config file into the initramfs, we need to support regenerating it
client side too.
Down the line, we'll need this to support overriding the kernel too.
This changes things in the core to add the concept of an "empty"
`RpmOstreeContext`. I initially tried skipping it, but that was too much
duplication. We still want all of the core ostree-related logic that lives in
that code too.
The treespec bits barfed if the spec didn't have a `tree/packages` key. It was
simplest to change that to allow it - and because that was the only case where
we errored out in parsing, I dropped the error handling.
There was another place in the upgrader that now needed to be fixed to handle
transitioning from just regenerating initramfs to not.
Closes: #574
Approved by: jlebon
I think since I landed a change to `--enable-new-name`, `/usr/bin/rpm-ostree`
became a symlink, and without the `-l` switch, rsync skips over them.
Hence, we have only been testing in vmcheck the old binaries, not new ones.
Oops.
Closes: #585
Approved by: jlebon
In the container build + vagrant test workflow, we have to run `make
vmsync` from the env in which autoconf was run, which is the container.
However, when using git worktrees (which is useful in this scenario to
avoid having to overwrite your local builds) and a CentOS build
container, the git inside is too old to understand worktrees, so let's
avoid invoking git in any of the test harness scripts.
Also make sure to restorecon after installing the new rpm-ostreed.
Closes: #555
Approved by: cgwalters
Add a few more tests to exercise some of the treefile options. We do
need to also expand test-basic.sh itself to sanity-check the structure
of a normal ostree compose. That's up next on the list.
Closes: #548
Approved by: cgwalters
This is just the final bit required to make sure the vagrant and
non-vagrant paths can work happily together. It's mostly minor fixes,
though the most major change which also affects vagrant is that we now
sync to the root home dir, rather than ~vagrant.
Closes: #524
Approved by: cgwalters
My development environment is now using "pet" docker containers.
I use VMs for testing things that require that (like rpm-ostree).
This patch builds on work from @jlebon in
https://github.com/projectatomic/rpm-ostree/pull/509
to rework `vmcheck` such that it can work on any `ssh-config`. By
default we expect this to be Vagrant.
However, I go a lot farther and delete the `vmbuild` code that was
trying to do builds in a container on the target VM. I think this is
still worth pursuing at some point, but for now I think it's
reasonable to assume that the rpm-ostree developer audience uses Linux
as their host workstation and hence has containers.
(There's another important point here in that for developing lower
level things like rpm-ostree, there's a strong push to make the VM
disposable and not a pet)
Closes: #516
Approved by: jlebon
Thought it'd be fun to write a test for verifying proper handling of
scriptlets during package layering. There's obviously a lot more that
could go in here (patches welcome!), but it's a start.
Closes: #434
Approved by: cgwalters
This is a useless test for now since we should be able to test on a VM
even if there are updates available.
I initially had it just as an example of the vmcheck test harness,
without actually thinking too much into what it was testing. But we do
need a nice test-basic.sh that will give rpm-ostree a good general
workout.
Closes: #394
Approved by: cgwalters
- Rename test-layering.sh to test-layering-basic.sh and make it test
both pkg-add and pkg-remove.
- Add test-layering-relayer.sh, which verifies that pkgs are properly
relayered during the creation of new deployments (e.g. upgrades,
rebases, deploys).
- Add test-layering-rpmdb.sh, which verifies that packages respect the
rpm requirements before being overlayed.
Closes: #371
Approved by: cgwalters
Print out a better summary that sums up the passes, failures, and skips.
Silence the restoring step at the end of the testsuite by only echo'ing
to the log file. Finally, add some colouring because colours are
awesome.
Closes: #371
Approved by: cgwalters
We now make the test harness handle restoring the VM to the original
state. The wonderful thing about ostree here is that it's a perfect
shoo-in for this. We make a 'backup' of the current ref, and just have
to make sure that the VM is back on that ref after running each test.
This will allow us to write tests without worrying as much about
cleaning up in the event of an error.
Closes: #360
Approved by: cgwalters
This is a preliminary package layering test which simply installs the
foo package and verifies that it functions properly. A bunch of
primitives are added to libvm.sh to facilitate this and future tests.
Closes: #344
Approved by: cgwalters
This is a very simple test harness inspired by the atomic one. It's a
simple bash script that sets up a permanent ssh connection to the host
and runs the test scripts. Also add a "demo" test-basic.sh test to make
sure that it works.
Closes: #344
Approved by: cgwalters
If the dir that was rsync'ed already contains build artifacts, we don't
want those contaminating our build process.
Closes: #336
Approved by: cgwalters
The `make vmshell` target makes it even easier to immediately try out
your changes in a live running Atomic Host. It will automatically
provision the VM, sync your latest changes, build, install in a new
deployment onto which the VM is rebooted, and drop you in the shell.
Closes: #321
Approved by: cgwalters
In order for vmcheck to be useful for testing, we need the machine to be
in a "clean" state. That is, sitting on a commit, ready for being
manipulated.
This is a small step towards this goal. Instead of overwriting the
files, we use rofiles-fuse to safely install and create a new
deployment.
Closes: #321
Approved by: cgwalters
I find myself not liking the name rpm-ostree anymore; it's
descriptive, but unfortunately we compete with other projects with
easier and sexier names.
Also, people continually find the ostree and rpm-ostree layering
unclear. It's *much* easier to say "nxs depends on ostree", even
though textually it's obvious "rpm-ostree depends on ostree".
Anyways, just an experiment for now.
Closes: #307
Approved by: giuseppe
This is infrastructure that's oriented around Vagrant, we do builds
inside the target VM (actually inside a Docker container), but then
directly `make install DESTDIR=/host/`.
The goal here is to have a convenient workflow for:
- dev -> hand testing
- dev -> run destructive tests
Closes: #303
Approved by: jlebon
