* fix typo?
Not sure if that's what was intended by the original author, but there are 1000000 microseconds in one second, not 10000.
* update example command as well
* Mac: fix name of the vm directory
Actually the directory name changed from vm to vms/0 (there has been
no release with vm, we may remove references to it). This is for the
forthcoming 18.01 Edge release, but since it's already published we
should probably update now.
Signed-off-by: Akim Demaille <akim.demaille@docker.com>
* Toolbox: formatting changes
No user visible changes here, just making future diffs easier to read.
Signed-off-by: Akim Demaille <akim.demaille@docker.com>
* Toolbox: do not invite to remove ~/.docker
Docker for Mac uses that directory too.
Signed-off-by: Akim Demaille <akim.demaille@docker.com>
* Mac: the disk image has moved
In the forthcoming release of Docker for Mac, the disk image is in
another directory (was needed for future changes).
Signed-off-by: Akim Demaille <akim.demaille@docker.com>
* Mac: FAQ: Promote "Disk Usage" questions
One level upper.
Signed-off-by: Akim Demaille <akim.demaille@docker.com>
* Updated the options docs for journald driver
I have updated the options documentation for the journald logging driver, since I was having trouble understanding how it was supposed to be used. I modeled it heavily after the google cloud logging driver docs, since I found those helpful in solving my own issues.
* Make 'use a process manager' clearer
The "Use a process manager" section appears to mix use of process manager on host with using process managers within a container.
The essence of the section appears to suggest that host-level process managers are ok as long as you don't combine them with docker restart policies. But paragraphs 2 & 3 (in the original) seem to actively discourage the use of process managers within containers. Putting paragraphs 2 & 3 under an explicit heading would be clearer.
Should a section on process mangers within containers be expanded? There are lots of articles on the internet that relate to using process managers to start multiple processor in containers or to manage zombie pid 1 processes. Are there official recommendations on what to do in these situations — whether to use a process manager or is a recommended 'docker native' way to achieve the same results?
* Updated to use the warning class
When running a container with tty enabled and setting the journald
logging driver, you may see "blob data" instead of the real log
message.
That's because tty appends "\r" to the end of the log line,
which isn't stripped by jorunald unless you specify the "--all" flag.
Signed-off-by: Boaz Shuster <ripcurld.github@gmail.com>
The paragraph mentions that live restore is only supported for minor
releases (aka patch releases) but goes on to show "1.12.1" updated to
"1.13.2". This is unfortunately a pretty nasty typo and should be
"1.12.2" not "1.13.2".
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
In the code sample mentioned in `Start a service with volumes` it says
> The following example starts a nginx service with four replicas, each of which uses a local volume called myvol2.
But --replicas=4 is missing from the code sample which reads:
```
$ docker service create -d \
--name devtest-service \
--mount source=myvol2,target=/app \
nginx:latest
```
Instead of this it rather should be:
```
$ docker service create -d \
--replicas=4 \
--name devtest-service \
--mount source=myvol2,target=/app \
nginx:latest
```
The symlinks are described the wrong way round and suggest that /dev/stdout and /dev/stderr are being overwritten, but that's not true - ln -s /dev/stdout /var/log/nginx/access.log creates a symlink at access.log pointing to /dev/stdout.
See `man ln`
