Fixes: #18872
For backup, restore case, we usually restore in same directory, so
when we untar it, we not need go to /dbdata directory, wich could
result in /dbdata/dbdata, this fix make it untar directly. The directory
/dbdata would be same as old container /dbdata.
Signed-off-by: Kai Qiang Wu(Kennan) <wkqwu@cn.ibm.com>
The client API at fsouza/go-dockerclient has dependencies on packages in
the docker/docker repository which currently do not build on Solaris. In
particular, stat_unsupported.go makes use of the Mtimespec field of the
syscall.Stat_t struct, which is not present on Solaris, and a number of
Unix-specific packages do not list Solaris in their compile targets.
This commit adds enough support to be able to build
fsouza/go-dockerclient on SmartOS using Go 1.5.1 without affecting other
platforms.
Signed-off-by: James Nugent <james@jen20.com>
- Stop serializing JSONMessage in favor of events.Message.
- Keep backwards compatibility with JSONMessage for container events.
Signed-off-by: David Calavera <david.calavera@gmail.com>
1. Introduce pending state. Pending nodes need validation before moving to healthy state. Resolve issues of duplicate ID and dead node drop issues.
2. Expose error and last update time in docker info.
3. Use connect success/failure to drive state transition between healthy and unhealthy.
Signed-off-by: Dong Chen <dongluo.chen@docker.com>
Markdown linter produced an error on this page;
running markdownlint
ERROR (registry/spec/manifest-v2-2.md) frontmatter: Unexpected non-whitespace char: # Image Manifest Version 2, Schema 2
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Adds the new label to the documentation. The
impact/distribution label is intended for changes
that affect the image-format or interaction
with the registry (distribution).
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Also, fix up some minor whitespace consistency issues, remove a little cruft, and update GOARM for armhf to 7 so that we're explicit.
Signed-off-by: Andrew "Tianon" Page <admwiggin@gmail.com>
Only two of these are not handled, the one in `TestBuildForceRm`
is intended to not be handled, while the other one in
`TestBuildResourceConstraintsAreUsed` causes problem.
In test case `TestBuildResourceConstraintsAreUsed`, somehow we
are not able to access network to get base image, but the error
message is:
```
Error: failed to inspect container : Unable to read inspect data: json: cannot unmarshal array into Go value of type types.ContainerJSON
```
Totally confusion.
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
GetCryptoService is a helper function and we have injected a specific
http.ResponseWriter object for it to write back error message. Meanwhile
the caller for that function checks whether the cryptoService is nil
or not and return immediately if it is nil.
I think it's not a good idea to write back HTTP response in the helper function,
it's the caller's work and thus there is no need to inject the specific
ResponseWriter object into it.
Signed-off-by: Hu Keping <hukeping@huawei.com>
When a container create with -m 100m and then docker update other
cgroup settings such as --cpu-quota, the memory limit show by
docker stats will become the default value but not the 100m.
Signed-off-by: Lei Jitang <leijitang@huawei.com>
- Use the ones provided by docker/go-connections, they are a drop in replacement.
- Remove pkg/sockets from docker.
- Keep pkg/tlsconfig because libnetwork still needs it and there is a
circular dependency issue.
Signed-off-by: David Calavera <david.calavera@gmail.com>
After running into a few nil pointer errors during development, it is clear
that having this function return nil when a hash is not available is the wrong
approach. Nearly every time, this lack of availability was due to a missing
import statement for the hash. This is always a programming error.
To avoid future confusion, we now appropriately panic when the hash function is
not imported by the application. More dynamic uses of the package should call
Algorithm.Available() before calling Algorithm.Hash() to avoid this panic.
Signed-off-by: Stephen J Day <stephen.day@docker.com>
