Since commit 925b484 ("No fallback nameservers for internal
resolver"), if the host's resolv.conf has no nameservers and
no servers are supplied via config, the internal resolver will
not use Google's DNS - so the container will not be able to
resolve external DNS requests.
That can happen when container's are "restart-always" and the
docker daemon starts before the host's DNS is configured.
So, to highlight the issue (which may not be an error, but
probably is), include a warning in the container's resolv.conf
file.
Also, log a warning - logs currently say "No non-localhost DNS
nameservers are left in resolv.conf. Using default external
servers". But, that's misleading because it's from an initial
resolv.conf setup, before the internal resolver configured without
those fallbacks - we'll drop the fallbacks completely once the
default bridge has an internal resolver).
Signed-off-by: Rob Murray <rob.murray@docker.com>
Since commit 51d7f95 ("libnet: remove struct endpointCnt") an
endpoint count for networks has not been persisted.
But, on downgrade to a version older than that commit, the
missing field caused daemon startup to fail.
So, create the count in the store - it only needs to exist, it's
no longer maintained as a count of endpoints. On downgrade, the
count is probably zero anyway (the daemon is stopped), but the
older daemon fixes it up on startup if necessary.
Signed-off-by: Rob Murray <rob.murray@docker.com>
The existing code was more complicated than needed. By default, the daemon
runs "unconfined", but we try to detect the current profile that's set.
When failing to do so (error, or detected profile is empty), we assume
the default ("unconfined").
This patch simplifies the logic;
- Set the default ("unconfined")
- Only update the default when we successfully found the current profile
(no error occurred, and the profile is not empty).
While updating, also;
- Replaced use of `strings.SplitN` for `strings.Cut`, which is more
efficient, and doesn't allocate.
- Move constructing the profileData closer to where it's used.
- Remove intermediate var.
- Combine defers and change the order (close file first, before removing),
and suppress errors to keep linters happy.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
`exec.Cmd.Path` already contains the command that was executed, so we
were printing the command twice. However, `exec.Cmd` implements a stringer
interface, which provides a readable version of the command that was
executed, so use that instead. While updating, lso change backticks in
the error for regular quotes.
Before:
running `/usr/sbin/apparmor_parser apparmor_parser -Kr /no/such/file` failed with output: Cache read/write disabled: interface file missing. (Kernel needs AppArmor 2.4 compatibility patch.)
Warning: unable to find a suitable fs in /proc/mounts, is it mounted?
Use --subdomainfs to override.
error: exit status 1
After:
running '/usr/sbin/apparmor_parser -Kr /no/such/file' failed with output: Cache read/write disabled: interface file missing. (Kernel needs AppArmor 2.4 compatibility patch.)
Warning: unable to find a suitable fs in /proc/mounts, is it mounted?
Use --subdomainfs to override.
error: exit status 1
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We are considering moving the apparmor profile to a separate module,
so removing these comments in preparation. These comments are ignored
already when building in go module mode, so have little benefits.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We are considering moving the seccomp profile to a separate module,
so removing these comments in preparation. These comments are ignored
already when building in go module mode, so have little benefits.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We are considering moving the seccomp profile to a separate module,
so reducing the list of dependencies for this package.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This reverts commit 5d2006256f, which
caused some issues in the docker/cli formatting code that needs some
investigating.
Let's (temporarily) revert this while we look what's wrong.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Copy the swagger / OpenAPI file to the documentation. This is the API
version used by the upcoming v28.1.0 release.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Like for containerCreate, filter the list of container labels based on
`DOCKER_OTEL_INCLUDE_CONTAINER_LABEL_ATTRS` and put that list in the
OTel span.
Signed-off-by: Albin Kerouanton <albinker@gmail.com>
This commit adds a new OTel span to `(*Daemon).containerCreate()` and
puts filtered container labels in the span attributes.
The filter is based on a comma-separated list of labels provided through
the `DOCKER_OTEL_INCLUDE_CONTAINER_LABEL_ATTRS` environment variable.
This label might be removed at any point in time if we figure out a
better way to filter labels, or if that span becomes unecessary.
Signed-off-by: Albin Kerouanton <albinker@gmail.com>
It's possible to remove a legacy link from running containers.
When that happens, the Sandbox's Endpoints are removed and
re-added.
Since commit 65120d5 ("Create bridge veth in container netns")
the veth device has been created in the container's netns. When
that happens, a flag is set on the Endpoint to note that it
does not need to be moved into the netns.
But, during the Leave/Join (Sandbox.Refresh) the veth is moved
out of the netns. So, clear the flag during the Leave, to note
that it needs to be moved back in during the Join.
Signed-off-by: Rob Murray <rob.murray@docker.com>
The `BridgeNfIptables` and `BridgeNfIp6tables` fields in the
`GET /info` response were deprecated in API v1.48, and are now omitted
in API v1.49.
With this patch, old API version continue to return the field:
curl -s --unix-socket /var/run/docker.sock http://localhost/v1.48/info | jq .BridgeNfIp6tables
false
curl -s --unix-socket /var/run/docker.sock http://localhost/v1.48/info | jq .BridgeNfIptables
false
Omitting the field in API v1.49 and above
curl -s --unix-socket /var/run/docker.sock http://localhost/v1.49/info | jq .BridgeNfIp6tables
null
curl -s --unix-socket /var/run/docker.sock http://localhost/v1.49/info | jq .BridgeNfIptables
null
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This was introduced in c817ea2159, but overlooked
that the logs were now produced before (for POST requests) the form-data was
handled and added to the "fields" variable.
This patch changes the logic to wrap the handler and make sure logs are
created before, and after the handler is called (the "form-data" is always
added before the handler is called).
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
