Multus requires that the 'ip' tool be installed in the Rook image.
Today, it is present in the upstream Ceph base image that is used, but
that is expected to change in the future. The Ceph project intends to
switch to CentOS 9 minimal images, and the 'ip' tool will likely be
removed from the base image at that point.
However, Rook CI has occasionally had issues with the current 'dnf
install' command when CentOS repos go down, or when there is otherwise
some problem. Because the package is already installed today, there is
no need to hamstring Rook builds when CentOS is having problems. But we
do want to make sure that Rook builds don't silently succeed in the
eventual future when 'ip' tool is removed from the Ceph image.
For now, replace the 'dnf install' with a check to verify that 'ip' tool
is installed, and add a shorter form of this note as a comment above it
to help Rook maintainers know how to resolve future 'ip' tool removal.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
(cherry picked from commit c4e99c1462)
The rook/ceph Dockerfile uses dnf to ensure iproute (containing the 'ip'
CLI tool) is installed in the Rook image for Multus usage. This comes
from the 'baseos' repo, but if any other repos are unavailable
temporarily, it can cause the container build to fail.
Use the '--repo baseos' flag to help prevent these kinds of failures.
Additionally, this will speed up the build slightly since it does not
attempt to load any non-necessary repos.
This change may make the container build slightly fragile in the future
if CentOS changes the name of its baseos repo, or if the Ceph image
switches to a non-CentOS base image.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
Use the Rook image (defined by the operator pod) to detect the Multus
network address ranges. It is reasonable for users to want to have a
minimal Ceph image that does not have the `ip` utility installed, which
is used for detecting the address ranges of multus interfaces. Instead,
use the Rook image, which Rook can ensure has the `ip` tool if Ceph ever
removes it from their image.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
The toolbox can now interact with S3 gateways using the `s5cmd` tool.
The binary is only 12M so this does not add up too much to the operator
image size.
Closes: https://github.com/rook/rook/issues/4968
Signed-off-by: Sébastien Han <seb@redhat.com>
The rook operator as well as the toolbox pod run with the "rook" user
with UID 2016. The UID was chosen based on the year of the initial
commit in the rook/rook repository.
No more root user running.
Closes: https://github.com/rook/rook/issues/8734
Signed-off-by: Sébastien Han <seb@redhat.com>
We don't need to use tini.
We don't have anything in the rook operator that would
either create zombie processes (no threads) or use
exec (to fork). The Go binary has a really good
signal handling mechanism.
Closes: https://github.com/rook/rook/issues/8794
Signed-off-by: Sébastien Han <seb@redhat.com>
The flex driver has been fully deprecated and thus removed from Rook.
Before upgrading to v1.8, users will need to convert existing flex volumes
from flex to csi volumes.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Thanks to Golang 1.16, we can now embed files in the Go binary. This
means we don't need to add the CSI templates files to the container
image. They are added in the Go binary at build time.
The existing location of the template must be in the package calling it.
So they moved to pkg/operator/ceph/csi/template. All the files have been
symlinked back to cluster/examples/kubernetes/ceph/csi/template.
Closes: https://github.com/rook/rook/issues/7609
Signed-off-by: Sébastien Han <seb@redhat.com>
From inside the toolbox, the ceph debug level can now be changed for all
ceph logging subsystems, for instance, level 1 will be:
```
kubectl -n rook-ceph exec deploy/rook-ceph-tools -- set-ceph-debug-level 1
```
To reset to default values just run:
```
kubectl -n rook-ceph exec deploy/rook-ceph-tools -- set-ceph-debug-level default
```
Signed-off-by: Sébastien Han <seb@redhat.com>
Include the mandatory template files in the rook ceph image
so that the user-driven configuration for CSI and flex do not
need to differ in this aspect.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Commit 91cc62676d changed our base ceph
version from Mimic to Nautilus.
Update the comment in the Dockerfile to reflect this change.
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Instead of adding Ceph packages to Rook's base image, use
ceph-container's base image, which has Ceph packages preinstalled. Rook
now benefits from the testing the ceph-container project does, and the
scope of the Rook project no longer has to maintain builds of base
images containing Ceph packages.
The only ARM builds of ceph-container are CentOS, and as switching to
CentOS was already a thought, that is easily done at the same time as
this change.
Signed-off-by: Blaine Gardner <blaine.gardner@suse.com>
Ceph is now installed from deb packages instead of being built from
source. This results in larger images sizes (roughly 60MB increase
compressed) but the maintenance and build burden is reduced. Future
versions of ceph will hopefully optimize for image size.
All support for building native code in containers has been removed,
as it is no longer needed. When we add other backends we will assume
that they can be installed using apt-get too.
Blaine Gardner
subhamkrai
Sébastien Han
Travis Nielsen
David Vossel
Umanga Chapagain
John Mulligan
Ken Dreyer
Fan, Shang Xiang
Blaine Gardner
Jared Watts
Bassam Tabbara
Alexander Trost
Bassam Tabbara