Hash the container ID, mount source and destination together to form a
layer name.
This ensures the generated names are filesystem-friendly and don't
exceed path length limits while maintaining uniqueness across different
mount points and containers.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
(cherry picked from commit cb88c6ba10)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Add cleanup for the init layer directory if any operation fails after
driver.CreateReadWrite() succeeds in initMount(). Previously, failures
in driver.Get(), initFunc(), or driver.Put() would leave an orphaned
overlay2 directory.
Related to moby/moby#45939
Signed-off-by: Jan Scheffler <jan.scheffler@qodev.ai>
(cherry picked from commit 3fdde529e7)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Add cleanup for the RW layer directory if saveMount() fails after
driver.CreateReadWrite() succeeds. Previously, this failure path would
leave an orphaned overlay2 directory with no corresponding metadata.
Related to moby/moby#45939
Signed-off-by: Jan Scheffler <jan.scheffler@qodev.ai>
(cherry picked from commit d7a6250b91)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Start the metadata transaction before creating the overlay2 directory.
This ensures that if driver.Create() fails, we can properly cancel the
transaction. Previously, if StartTransaction() failed after driver.Create()
succeeded, the defer cleanup would not run (not registered yet), leaving
an orphaned overlay2 directory.
The fix reorders operations so that:
1. Transaction is started first (no filesystem changes yet)
2. Overlay2 directory is created second (transaction ready for cleanup)
3. Defer is registered after both succeed (tx is guaranteed non-nil)
If driver.Create() fails, the transaction is explicitly cancelled before
returning. The nil check for tx in the defer is no longer needed since
tx is guaranteed to exist when the defer runs.
Related to moby/moby#45939
Signed-off-by: Jan Scheffler <jan.scheffler@qodev.ai>
(cherry picked from commit 70004549fb)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Stopping the Engine while a container with autoremove set is running may
leave behind dead containers on disk. These containers aren't reclaimed
on next start, appear as "dead" in `docker ps -a` and can't be
inspected or removed by the user.
This bug has existed since a long time but became user visible with
9f5f4f5a42. Prior to that commit,
containers with no rwlayer weren't added to the in-memory viewdb, so
they weren't visible in `docker ps -a`. However, some dangling files
would still live on disk (e.g. folder in /var/lib/docker/containers,
mount points, etc).
The underlying issue is that when the daemon stops, it tries to stop all
running containers and then closes the containerd client. This leaves a
small window of time where the Engine might receive 'task stop' events
from containerd, and trigger autoremove. If the containerd client is
closed in parallel, the Engine is unable to complete the removal,
leaving the container in 'dead' state. In such case, the Engine logs the
following error:
cannot remove container "bcbc98b4f5c2b072eb3c4ca673fa1c222d2a8af00bf58eae0f37085b9724ea46": Canceled: grpc: the client connection is closing: context canceled
Solving the underlying issue would require complex changes to the
shutdown sequence. Moreover, the same issue could also happen if the
daemon crashes while it deletes a container. Thus, add a cleanup step
on daemon startup to remove these dead containers.
Signed-off-by: Albin Kerouanton <albin.kerouanton@docker.com>
(cherry picked from commit ec9315cd4f)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
This code was added in 85b260fba8, but didn't
account for maps.Clone returning a `nil` map if the map cloned was `nil`.
This could lead to a panic, similar to the panic that was fixed in
7517464283d29969c4d3615397b369abd99ce395;
panic: assignment to entry in nil map
goroutine 498 [running]:
github.com/moby/moby/v2/daemon.buildPortsRelatedCreateEndpointOptions(0x400042f348, 0xaaaabcc8f458?, 0x40006feb40)
/root/build-deb/engine/daemon/network.go:1047 +0x844
github.com/moby/moby/v2/daemon.buildCreateEndpointOptions(0x400042f348, 0x4001015040, 0x400027d320, 0x40006feb40, {0x0, 0x0, 0x4001506cb8?})
/root/build-deb/engine/daemon/network.go:988 +0x20c
github.com/moby/moby/v2/daemon.(*Daemon).connectToNetwork(0x4000898008, {0xaaaabe21d9f8, 0x4000f12b10}, 0x400089a008, 0x400042f348, {0x400077a9f0, 0x6}, 0x400027d320)
/root/build-deb/engine/daemon/container_operations.go:738 +0x66c
github.com/moby/moby/v2/daemon.(*Daemon).allocateNetwork(0x4000898008, {0xaaaabe21d9f8, 0x4000f12b10}, 0x400089a008, 0x400042f348)
/root/build-deb/engine/daemon/container_operations.go:421 +0x298
github.com/moby/moby/v2/daemon.(*Daemon).initializeCreatedTask(0x4000898008, {0xaaaabe21d9f8, 0x4000f12b10}, 0x400089a008, {0xaaaabe23dc60, 0x4000eb21c8}, 0x400042f348, 0xaaaabd4db3df?)
/root/build-deb/engine/daemon/start_linux.go:37 +0x260
github.com/moby/moby/v2/daemon.(*Daemon).containerStart(0x4000898008, {0xaaaabe21d9c0, 0xaaaabfa05300}, 0x400089a008, 0x400042f348, {0x0, 0x0}, {0x0, 0x0}, 0x1)
/root/build-deb/engine/daemon/start.go:242 +0xba8
github.com/moby/moby/v2/daemon.(*Daemon).restore.func4(0x400042f348, 0x400117f1f0)
/root/build-deb/engine/daemon/daemon.go:633 +0x308
created by github.com/moby/moby/v2/daemon.(*Daemon).restore in goroutine 1
/root/build-deb/engine/daemon/daemon.go:607 +0x5ec
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 695010ba2e)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
If a container is using a docker_gwbridge endpoint as its gateway,
when it's connected to another network that provides a gateway, the
docker_gwbridge endpoint is removed when that endpoint is added (in
a recursive nightmare).
So, the "before" gateway for the container has been removed
before the new gateway is updateExternalConnectivity'd.
Don't pass the old gateway to updateExternalConnectivity in that
case, because the network driver's already forgotten about it.
Signed-off-by: Rob Murray <rob.murray@docker.com>
(cherry picked from commit 1731e9e729)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We've seen various failures recently where GitHub actions runners are
running out of space. Skip this test for now.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 3e4a3cb03e)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The priority order for determining image store choice was incorrect when
a prior graphdriver existed.
The issue occurred because the prior graphdriver check happened after
processing explicit driver configuration, effectively ignoring user
intent when prior state existed.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
(cherry picked from commit 391247ce96)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
If the DNS name still resolves to an IP address, and that address is
assigned to a running container, the ping command will run indefinitely
and the test suite will time out for 10 mins.
This is confusing, as it looks like a daemon hang, or a test suite hang,
whereas it's just a test failure. Add '-c1' to ping to make it return
immediately.
Signed-off-by: Albin Kerouanton <albin.kerouanton@docker.com>
(cherry picked from commit 53ea70ea46)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Previous commit reverted a faulty change that broke DNS resolution for
non swarm-scoped networks once a node has joined a Swarm cluster.
This commit adds an integration test to verify that we don't break DNS
resolution again.
Signed-off-by: Albin Kerouanton <albin.kerouanton@docker.com>
(cherry picked from commit 47bd247d4d)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Commit a8b9eff90 removed a call to Network.updateSvcRecord from
Network.createEndpoint on the grounds that:
> all callers of Network.createEndpoint follow up with an Endpoint.Join,
> which also sets up the DNS entry.
However, the original call in Network.createEndpoint was gated by:
```
if !n.getController().isSwarmNode() || n.Scope() != scope.Swarm || !n.driverIsMultihost() {
n.updateSvcRecord(context.WithoutCancel(ctx), ep, true)
}
```
whereas the call in Endpoint.sbJoin() (invoked by Endpoint.Join()) is
gated by:
```
if !n.getController().isAgent() {
if !n.getController().isSwarmNode() || n.Scope() != scope.Swarm || !n.driverIsMultihost() {
n.updateSvcRecord(context.WithoutCancel(ctx), ep, true)
}
}
```
As a result, once a node has joined a Swarm cluster, no DNS entries are
created for non swarm-scoped networks.
Change the condition used by `sbJoin` to match the original condition
used in `createEndpoint`.
Signed-off-by: Albin Kerouanton <albin.kerouanton@docker.com>
(cherry picked from commit 2e41476a5f)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
The image exporter wrapper was unconditionally setting `unpack=true` for
all build exports, preventing users from controlling this behavior
through buildkit's output image exporter option.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
(cherry picked from commit b4f9bd1cb3)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Running sub-package tests from the root module without readding the
replace rules wasn't running the tests from the local in-tree versions
of these submodules.
Fix by cd-ing into their directories before running tests.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
(cherry picked from commit 0f597561e8)
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
2025-11-13 20:42:58 +01:00
29 changed files with 197 additions and 90 deletions
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
