Run `xfs_repair` on XFS filesystems that needs repairing indicated by
the `unix.EUCLEAN` error when mounting
Fixes#5319Fixes#5437
Signed-off-by: Noel Georgi <git@frezbo.dev>
Fixes#4694
User services run alongside with Talos system services.
Every user service container root filesystem should be already present
in the Talos root filesystem.
Signed-off-by: Andrey Smirnov <andrey.smirnov@talos-systems.com>
The list of layers should come from the `/extensions.yaml` configuration
file.
Closes: https://github.com/talos-systems/talos/issues/4814
Signed-off-by: Artem Chernyshev <artem.chernyshev@talos-systems.com>
If the mount is skipped, we shouldn't record it and create a matching
resource.
This fixes a problem discovered by cluster discovery tests when node
establishes more than a single identity on initial boot with first one
being lost, but still exists in the discovery service.
Signed-off-by: Andrey Smirnov <andrey.smirnov@talos-systems.com>
Initialize state should be only reading the config.
So now if it detects that the partition is not even formatted it will
skip it and will consider the state to be empty.
Signed-off-by: Artem Chernyshev <artem.0xD2@gmail.com>
The structure of the controllers is really similar to addresses and
routes:
* `LinkSpec` resource describes desired link state
* `LinkConfig` controller generates `LinkSpecs` based on machine
configuration and kernel cmdline
* `LinkMerge` controller merges multiple configuration sources into a
single `LinkSpec` paying attention to the config layer priority
* `LinkSpec` controller applies the specs to the kernel state
Controller `LinkStatus` (which was implemented before) watches the
kernel state and publishes current link status.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This fixes a race condition between `udevd` issuing ioctl `BLKRRPART`
when block device is closed after partitioning/formatting and Talos
trying to mount a partition. When `BLKRRPART` is issued, kernel
temporarily wipes out all the in-memory partitions killing `/dev/sdX`
devices until partition scan is done.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
Filesystem creation step is moved on the later stage: when Talos mounts
the partition for the first time.
Now it checks if the partition doesn't have any filesystem and formats
it right before mounting.
Additionally refactored mount options a bit:
- replaced separate options with a set of binary flags.
- implemented pre-mount and post-unmount hooks.
And fixed typos in couple of places and increased timeout for `apid ready`.
Signed-off-by: Artem Chernyshev <artem.0xD2@gmail.com>
This fixes spurious race conditions when user disks are partitioned
and formatted in `mountUserDisks` task. While this task runs, `udevd` is
running to allow various `/dev/` symlinks to be used for user disks.
At the same time `udevd` might trigger syscall `BLKRRPART` at any time
concurrently with Talos which leads to a race on kernel side when Talos
tries to update kernel partition table while kernel does it on its own
as a result of `udevd` call.
As part of the fix, `RereadPartitionTable()` calls were removed (they
trigger `BLKRRPART` and they're not needed as Talos updates partition
table on its own).
Some cleanups to make sure blockdevice is open/closed just in matching
pairs (no lingering open blockdevice instances). This is import for
`WithExclusiveLock()` calls, as it would lead to a deadlock if previous
blockdevice instance is not closed.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This introduces the notion of a "board" in Talos. A board is an interface that is capable
of modifying the installation in specific ways for a given SBC. This also adds support for the
libretech_all_h3_cc_h5.
Signed-off-by: Andrew Rynhard <andrew@rynhard.io>
Fixes were applied automatically.
Import ordering might be questionable, but it's strict:
* stdlib
* other packages
* same package imports
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This skips writing partition table if partition doesn't have to be
resized (already resized or max size from the beginning).
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
For 0.6 -> 0.7 upgrade, in any case config.yaml is preserved and moved
from `/boot` to `/system/state`.
For single node upgrade, `EPHEMERAL` partition is not touched and other
partitions are re-created as needed.
Bump provision tests to 0.6/0.7 upgrades as we get closer to the new
release.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This fixes a bug introduced in #1982, the intention was to ignore
`EINVAL` on `unmount` when partition is no longer mounted, but the
change was wrong as it affected both `mount` and `unmount` code paths.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This unifies more code paths under the control of `install.Manifest` vs.
being split across the installer and manifest code.
There should be no functional changes now.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
Library `blockdevice` was extracted as `talos-systems/go-blockdevice`,
this PR finalizes the move by removing Talos copy of it.
Some functions around `mkfs`/`growfs` were extracted as `makefs`
package, as they depend on `cmd` package.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
Adds the ability to apply (replace) an existing node configuration with
a new one via the Machine API.
Fixes#2345
Signed-off-by: Seán C McCord <ulexus@gmail.com>
This moves `pkg/config`, `pkg/client` and `pkg/constants`
under `pkg/machinery` umbrella.
And `pkg/machinery` is published as Go module inside Talos repository.
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
Fixes#2272
`gofumpt` is now included into `golangci-lint`, but not the
`gofumports`, so we keep it using it as separate binary, but we keep
versions in sync with `golangci-lint`.
This contains fixes from:
* `gofumpt` (automated, mostly around octal constants)
* `exhaustive` in `switch` statements
* `noctx` (adding context with default timeout to http requests)
Signed-off-by: Andrey Smirnov <smirnov.andrey@gmail.com>
This adds the `/system` directory to provide a dedicated
directory for all system related runtime files.
Signed-off-by: Andrew Rynhard <andrew@rynhard.io>
This adds a sentinel error for a missing partition table. This error
is used to detect if a partition table already exists when setting
up user defined disks.
In addition to the fix, this removes a legacy parameter from the
`PartitionTable` method that indicated that the partition table
should be read. It is safer to just read it every time. Also, I
can't think of a case when the block device partition table is nil
and we want to read.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This is a rewrite of machined. It addresses some of the limitations and
complexity in the implementation. This introduces the idea of a
controller. A controller is responsible for managing the runtime, the
sequencer, and a new state type introduced in this PR.
A few highlights are:
- no more event bus
- functional approach to tasks (no more types defined for each task)
- the task function definition now offers a lot more context, like
access to raw API requests, the current sequence, a logger, the new
state interface, and the runtime interface.
- no more panics to handle reboots
- additional initialize and reboot sequences
- graceful gRPC server shutdown on critical errors
- config is now stored at install time to avoid having to download it at
install time and at boot time
- upgrades now use the local config instead of downloading it
- the upgrade API's preserve option takes precedence over the config's
install force option
Additionally, this pulls various packes in under machined to make the
code easier to navigate.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This adds unix.MS_REC to the shared mounts. We haven't seen any reports
of bugs yet, but in some testing I found that `Bidirectional` mounts don't
work unless the mount point is rshared.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This removes the github.com/pkg/errors package in favor of the official
error wrapping in go 1.13.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This package provides a consistent way for us to retry arbitrary logic.
It provides the following backoff algorithms:
- exponential
- linear
- constant
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
The gofumports does everything that gofumpt does with the addition of
formatting imports. This change proposes the use of the `-local` flag so
that we can have imports separated in the following order:
- standard library
- third party
- Talos specific
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
The gofumpt linter is a stricter drop-in replacement for gofmt. The
rules are ones that I strongly agree with and I think it would be better
if we added this linter instead of nit picking every PR.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This adds a well defined task for handling all overlay mount points that
are required by the system.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This moves to using BLKPG ioctl instead of BLKRRPART. BLKRRPART is older
and more sensitive to EBUSY errors. BLKPG has the potential to minimize
the changes of encountering an EBUSY error when manipulating partition
tables.
In looking at a comparison between BLKPG and BLKRRPART, it seems that
both have their pros and cons. Eventually a combination of the two may
serve us better, but for now I think BLKPG will get us further.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This changes the data partition name to something more appropriate. We
chose ephemeral to make it very clear that the disk should not be used
for application data.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This change aims to make installations more unified and reliable. It
introduces the concept of a mountpoint manager that is capable of
mounting, unmounting, and moving a set of mountpoints in the correct
order.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
This uses the correct mount flag for read-only.
We mistakenly had the flag for opening a file as read-only.
Signed-off-by: Andrew Rynhard <andrew@andrewrynhard.com>
