Container images can be created without passwd or group file, currently
if one of these containers gets run with a --user flag the container blows
up complaining about t a missing /etc/passwd file.
We just need to check if the error on read is ENOEXIST then allow the
read to return, not fail.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
When we read the conmon error status file, if Atoi fails to parse
the string we read from the file as an int, print the string as
part of the error message so we know what might have gone wrong.
Signed-off-by: Matthew Heon <mheon@redhat.com>
Instead of running a full sync after starting a container to pick
up its PID, grab it from Conmon instead.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
When we scan a container in runc and see that it no longer
exists, we already set ContainerStatusExited to indicate that it
no longer exists in runc. Now, also set an exit code and exit
time, so PS output will make some sense.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
When syncing container state, we normally call out to runc to see
the container's status. This does have significant performance
implications, though, and we've seen issues with large amounts of
runc processes being spawned.
This patch attempts to use stat calls on the container exit file
created by Conmon instead to sync state. This massively decreases
the cost of calling updateContainer (it has gone from an
almost-unconditional fork/exec of runc to a single stat call that
can be avoided in most states).
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
After stopping containers, we run updateContainerStatus to sync
our state with runc (pick up exit code, for example). Then we
proceed to not save this to the database, requiring us to grab it
again on the next sync. This should remove the need to read the
exit file more than once.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
it is not writeable by non-root users so there is no point in having
access to it from a container.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
For pods using cgroupfs, we were seeing some error messages in CI
from an inability to remove the pod CGroup, which was traced down
to the conmon cgroup still being present as a child. Try to
remove these error messages and ensure successful CGroup deletion
by removing the conmon CGroup first, then the pod cgroup.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
If for any reason slirp4netns fails at startup, podman waits
indefinitely. Check every second if the process is still running so
that we avoid to hang.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
It was setting the wrong variable (CamelCase)
in the wrong module ("main", not "libpod")...
Signed-off-by: Anders F Björklund <anders.f.bjorklund@gmail.com>
conmon creates a symlink to avoid using a too long UNIX path.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1641800
There is still one issue when the path length of the symlink has the
same length of the attach socket parent directory since conmon fails
to create the symlink, but that must be addressed in conmon first.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Only return `ErrCtrStateInvalid` errors when the mount counter is equal
to 1. Also fix the "can't unmount [...] last mount[..]" error which
hasn't been returned when the error passed to `errors.Errorf()` is nil.
Fixes: #1695
Signed-off-by: Valentin Rothberg <vrothberg@suse.com>
for the purposes of performance and security, we use securejoin to contstruct
the root fs's path so that symlinks are what they appear to be and no pointing
to something naughty.
then instead of chrooting to parse /etc/passwd|/etc/group, we now use the runc user/group
methods which saves us quite a bit of performance.
Signed-off-by: baude <bbaude@redhat.com>
We implement the securejoin method to make sure the paths to /etc/passwd and
/etc/group are not symlinks to something naughty or outside the container
image. And then instead of actually chrooting, we use the runc functions to
get information about a user. The net result is increased security and
a a performance gain from 41ms to 100us.
Signed-off-by: baude <bbaude@redhat.com>
run prepare() -- which consists of creating a network namespace and
mounting the container image is now run in parallel. This saves 25-40ms.
Signed-off-by: baude <bbaude@redhat.com>
prevent opening the same file twice, since we re-exec podman in
rootless mode. While at it, also solve a possible race between the
check for the file and writing to it. Another process could have
created the file in the meanwhile and we would just end up overwriting
it.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
when reading the output from conmon using the JSON methods, it appears that
JSON marshalling is higher in pprof than it really is because the pipe is
"waiting" for a response. this gives us a clearer look at the real CPU/time
consumers.
Signed-off-by: baude <bbaude@redhat.com>
We probably won't be able to initialize a firewall plugin when we
are not running as root, so we shouldn't even try. Replace the
less-effect EUID check with the rootless package's better check
to make sure we don't accidentally set up the firewall in these
cases.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
The newly introduced CRIU version check is now used to make sure
checkpointing and restoring is only used if the CRIU version is new
enough.
Signed-off-by: Adrian Reber <areber@redhat.com>
I've seen a runc zombie process hanging around, it is caused by not
cleaning up the "$OCI status" process. Also adjust another location
that has the same issue.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
We already have functions for retrieving the container's CGroup
path, so use them instead of manually generating a path.
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
We are still requiring oci-systemd-hook to be installed in order to run
systemd within a container. This patch properly mounts
/sys/fs/cgroup/systemd/libpod_parent/libpod-UUID on /sys/fs/cgroup/systemd inside of container.
Since we need the UUID of the container, we needed to move Systemd to be a config option of the
container.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
