Merge mantle into coreos-assembler #1152
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
podman: Only run podman network tests on fcos, use --net=none some
A minor fix for non x86 platforms to consider qemu-unpriv when getting the architecture in order to retrieve the correct kolet binary Closes: coreos#1074
Check for qemu-unpriv platform when converting board to arch
When using `-k` on `kola spawn`, it would seed the userdata with a v2 Ignition config. This of course doesn't work on FCOS. Detect the version instead and seed according to that.
kola/spawn: fix using v2 for FCOS when using -k
When getting the qemu port, the list of connections in /proc/net/tcp is checked and the localhost address is compares against. Ensure that the correct value is used for big and little endian arches
To ensure that we don't leak the `qemu` process by accident, just always set the `PDEATHSIG` attribute on it before spawning. That way, even if `kola` is abruptly killed, `qemu` will also be killed. The only exception is `kola spawn --remove=false`, where do *do* want to leak the `qemu` process. Pass that down as a flag. This also essentially enables support for the `--qemu-options` flag on the `unprivqemu` platform.
Allow passing an fd to `kola spawn` which scripts can use to easily retrieve information about the spawned hosts. I'm planning to make use of this in rpm-ostree's CI for example.
This switch will basically tell `kola spawn` that we're not interested in a shell, but we want it to hang around. In combination with `--json-info-fd`, this should allow scripts to leverage the CLI for easily standing up temporary VMs.
Check endianness when getting the qemu port
gcp: Add UEFI_COMPATIBLE and SECURE_BOOT
platform/unprivqemu: Drop restrict=yes
Some `kola spawn` enhancements
I plan to use this for cosa to make it a bit easier to log in manually to instances created by `cosa kola`. It will be set to a predictable path rather than a dir in `/tmp` that one may have to hunt down.
Adds the aliyun platform API; initially only supports a select set of ECS & OSS calls relating to importing & copying images.
I thought this passed at one time, but it seems to just hang here. I tried to debug it a little, but the expected flow here is opaque to me, and it just doesn't seem worth it. Bigger picture anyways, rather than inventing our own tests for most upstream projects, we should run *their* tests that have already been run in *their* CI too. This is what I plan to do for ostree/rpm-ostree at least with FCOS.
Now that CL is branched, let's improve ergonomics here and default to `fcos` for distro. This has bit me a few times already where I ran `kola` directly and forgot the `-b` argument.
kola: Change distro default to fcos, drop cl
kola: Remove podman network-multi test
kola/podman: Bump memory tests to 128m for runc
Adds a new aliyun command and 3 new subcommands underneath it all relating to image creation, copying, and deletion.
*: add aliyun
network: Honor MANTLE_SSH_DIR
Adds a test that verifies pods continue when the crio service is restarted. See: https://access.redhat.com/solutions/3825511 Signed-off-by: Steve Milner <[email protected]>
We just lifted the `restrict=yes` from the qemu-unpriv path which fixed the podman tests, but didn't remove it from the `PlatformsNoInternet` list. This meant that we were still skipping tests like `crio.base` that now *should* run with qemu-unpriv. (In the PR we discussed reworking all of this so that "no networking" is an opt-in thing for qemu-unpriv, which I'll do once we've drained the queue of other mantle PRs)
kola: Note qemu-unpriv now has networking
These only apply to `cl` which is on a branch now.
cgwalters
force-pushed
the
merge-mantle
branch
2 times, most recently
from
b7b7ae5 to
26956cc
Compare
cgwalters
changed the title
Regression from 7383128 which broke `coreos.ignition.mount.disks` because we were overriding the provided serial.
Changed from by-partlabel to by-partuuid by assigning a random uuid. On s390x, the by-partlabel string is incorrect because GPT partitioning uses little endian by default and s390x is big endian. Refer to: coreos/ignition-dracut#100.
qemu: Only set serial= if not already provided
kola: Use uuid instead of partlabel for ignition.mount disk test
cgwalters
force-pushed
the
merge-mantle
branch
from
26956cc to
5d8cee0
Compare
Regression from d4b252b We need to keep parsing the cosa build, just not actually use it to sync the target options in order to match how the upgrade test works.
kola: Fix syncOptions
Basically it doesn't make sense to separate building, testing, and uploads. There are too many entangled problem domains, among them: - How to run qemu - Parsing build schemas - Uploading to one AWS region, running tests, then replicating etc. This prepares the mantle project for merge into coreos-assembler by moving it to a subdirectory.
We are going to merge mantle into the project.
Basically it doesn't make sense to separate building, testing,
and uploads. There are too many entangled problem domains,
among them:
How to run qemu
Parsing build schemas
Uploading to one AWS region, running tests, then replicating
etc.
This merges the https://github.com/coreos/mantle project into
coreos-assembler.
Closes: coreos#163
Fixup for removing mantle git submodule.
cgwalters
force-pushed
the
merge-mantle
branch
from
d5ebdf8 to
bfcf0c3
Compare
|
OK, this should be good to go now! I had to enable merge commits on this repository rather than "rebase and merge". And I'm not sure Prow can merge it because of that, I think we have Prow set up to rebase. So this one will have to be someone clicking the button manually. |
|
/approve @cgwalters WELL DONE! |
|
+1,020,323... wonderful world of golang vendoring, here we come! |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: cgwalters, darkmuggle, jlebon The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
Note you will likely have to |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Basically it doesn't make sense to separate building, testing,
and uploads. There are too many entangled problem domains,
among them:
etc.
Closes: #163