The checksum is now being stored in a separate file beside the image
JSON file.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Fixes#10387
Without TCP keep-alive set on socket connections to the daemon, any
long-running container with std{out,err,in} attached that doesn't
read/write for a minute or longer will end in ECONNTIMEDOUT (depending
on network settings/OS defaults, etc.), leaving the docker client side
believing it is still waiting on data with no actual underlying socket
connection.
This patch turns on TCP keep-alive for the underlying TCP connection
for both TLS and standard HTTP hijacked daemon connections from the
docker client, with a keep-alive timeout of 30 seconds.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com>
This changeset addresses intermittent internal server errors encountered during
pushes. The root cause has been isolated to layers that result in identical,
empty filesystems but may have some path declarations (imaginge "./"),
resulting in different tarsums. The main error message reported during these
upload problems was a 500 error, which was not correct. Further investigation
showed the errors to be rooted in digest verification when finishing uploads.
Inspection of the surrounding code also identified a few issues. PutLayerChunk
was slightly refactered into PutLayerUploadComplete. Helper methods were
avoided to make handler less confusing. This simplification leveraged an
earlier change in the spec that moved non-complete chunk uploads to the PATCH
method. Simple logging was also added in the unknown error case that should
help to avoid mysterious 500 errors in the future.
At the same time, the glaring omission of a proper layer upload cancel method
was rectified. This has been added in this change so it is not missed in the
future.
In the future, we may want to refactor the handler code to be more
straightforward, hopefully letting us avoid these problems in the future.
Added test cases that reproduce these errors and drove these changes include
the following:
1. Push a layer with an empty body results in invalid blob upload.
2. Push a layer with a different tarsum (in this case, empty tar)
3. Deleting a layer upload works.
4. Getting status on a deleted layer upload returns 404.
Common functionality was grouped into shared functions to remove repitition.
The API tests will still require future love.
Signed-off-by: Stephen J Day <stephen.day@docker.com>
We've added support to the registry command to report the current version of
the distribution package. The version package is generated with a shell script
that gets the latest tag and add "+unknown". This allows builds from "go get"
and "go install" to have a rough version number. Generated periodically, it
will provide a decent indication of what code built the binary. For more
accurate versioning, one can build with the "binaries" make target. Linker
flags are used to replace the version string with the actual current tag at
build time.
Signed-off-by: Stephen J Day <stephen.day@docker.com>
While checksums are verified when a layer is pulled from v2 registries,
there are known issues where the checksum may change when the layer diff
is computed again. To avoid these issues, the checksum should no longer
be computed and stored until after it has been extracted to the docker
storage driver. The checksums are instead computed lazily before they
are pushed to a v2 registry.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Don't run docker in interactive mode in the archlinux container build script. Recent versions of Docker error out with "cannot enable tty mode on non tty input" when being run non-interactively (such as in a cron job, Jenkins build, etc.)
Signed-off-by: Jared Biel <jared.biel@bolderthinking.com>
