The `createAndCopy` function is only called when `nw.w` is nil
in order to create a new writer and prepare it. The current code
is attempting to close `nw.w` when there is a copy error. The
correct behavior would be to close the new writer and not touch `nw.w`.
Signed-off-by: Derek McGowan <derek@mcg.dev>
Bundle directory permissions should be 0700 by default. On Linux with
user namespaces enabled, the remapped root also needs access to the
bundle directory. In this case, the bundle directory is modified to
0710 and group ownership is changed to the remapped root group.
Port of the same change for the v2 runtime
Signed-off-by: Samuel Karp <skarp@amazon.com>
Bundle directory permissions should be 0700 by default. On Linux with
user namespaces enabled, the remapped root also needs access to the
bundle directory. In this case, the bundle directory is modified to
0710 and group ownership is changed to the remapped root group.
Signed-off-by: Samuel Karp <skarp@amazon.com>
This test will make sure there aren't any issues with multilayered
images during import. Keep in mind that in the case of multilayered
images, they have to be unpacked first in order to be usable.
Signed-off-by: Claudiu Belu <cbelu@cloudbasesolutions.com>
In linux platform, the shim server always listens on the socket before
the containerd task manager dial it. It is unlikely that containerd task
manager should handle reconnect because the shim can't restart. For this
case, the containerd task manager should fail fast if there is ENOENT or
ECONNREFUSED error.
And if the socket file is deleted during cleanup the exited task, it
maybe cause that containerd task manager takes long time to reload the
dead shim. For that task.v2 manager, the race case is like:
```
TaskService.Delete
TaskManager.Delete(runtime/v2/manager.go)
shim.delete(runtime/v2/shim.go)
shimv2api.Shutdown(runtime/v2/task/shim.pb.go)
<- containerd has been killed or restarted somehow
bundle.Delete
```
The shimv2api.Shutdown will cause that the shim deletes socket file
(containerd-shim-runc-v2 does). But the bundle is still there. During
reloading, the containerd will wait for the socket file appears again
in 100 seconds. It is not reasonable. The Reconnect should prevent this
case by fast fail.
Closes: #5648.
Fixes: #5597.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
This adds valuable logging data to the open telemetry traces.
When the trace is not recording we don't bother doing anything as it is
relatively expensive to convert logrus data to otel just due to the
nature of how logrus works.
The way this works is that we now set a context on the logrus.Entry that
gets passed around which the hook then uses to determine if there is an
active span to forward the logs to.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
On FreeBSD + zfs, stat call seem to return garbage in RDev for regular
files & folders. The value returned is large enough not to fit into
`Devmajor` & `Devminor` fields of the tar header. Fortunately, these
fields are required just for special devices.
This change
* adds a check into `setHeaderForSpecialDevice` that the
input header represents a special device. If it's not the case, we
don't set the Devmajor & Devminor fields.
* fixes the nil check on `getxattr`: it never returns nils, but rather
an empty slice instead
Signed-off-by: Artem Khramov <akhramov@pm.me>
The io/ioutil package has been deprecated as of Go 1.16, see
https://golang.org/doc/go1.16#ioutil. This commit replaces the existing
io/ioutil functions with their new definitions in io and os packages.
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
The `oci.WithUser` function relies on checking a path on the hosts disk to
grab/validate the uid:gid pair for the user string provided. For LCOW it's a
bit harder to confirm that the user actually exists on the host as a rootfs isn't
mounted on the host and shared into the guest, but rather the rootfs is constructed
entirely in the guest itself. To accomodate this, a spot to place the user string
provided by a client as-is is needed.
The `Username` field on the runtime spec is marked by Platform as only for Windows,
and in this case it *is* being set on a Windows host at least, but will be used as a
temporary holding spot until the guest can use the string to perform these same
operations to grab the uid:gid inside.
Signed-off-by: Daniel Canter <dcanter@microsoft.com>
Some cases can cause the server initialization to block (namely running
a 2nd containerd instance by accident against the same root dir). In
this case there is no way to quit the daemon except with `kill -9`.
This changes context things so that server init is done in a goroutine
and we wait on a channel for it to be ready while we also wait for a
ctx.Done(), which will be cancelled if there is a termination signal.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
clone3 is explicitly requested to give ENOSYS instead of the default EPERM, when CAP_SYS_ADMIN is unset.
See moby/moby PR 42681 (thanks to berrange).
Without this commit, rawhide image does not work:
```console
$ sudo ctr run --rm --net-host --seccomp registry.fedoraproject.org/fedora:rawhide foo /usr/bin/curl google.com
curl: (6) getaddrinfo() thread failed to start
```
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
During import, if an image does not match the host's platform,
it won't have any children labels set, which will result in the
Garbage Collector deleting its content later, resulting in an
unusable image. In this case, we should fail early.
This can still be bypassed by using ctr import --all-platforms.
Signed-off-by: Claudiu Belu <cbelu@cloudbasesolutions.com>
