feat(daemon) `ipfs daemon --initialized` #355

btc · 2014-11-16T21:44:37Z

With this PR, users can fetch and run ipfs in one command.

docker run -d -p 4001:4001 -p 5001:5001 jbenet/go-ipfs daemon --initialized

Magic.

nb: needs ipfs/go-datastore#9

whyrusleeping · 2014-11-17T04:12:48Z

👍 ❤️

whyrusleeping · 2014-11-17T04:15:49Z

is this PR accidentally mixed with your other logging PR?

btc · 2014-11-17T08:14:29Z

@whyrusleeping Good observation. No, it's not an accident.

Out of productivity concerns, both for me and the reviewers involved, I'm experimenting with a more fluid approach to PRs. I intend to include minor, general improvements--so long as they don't significantly harm readability or detract from the focus of the change set. Pulling out weeds as I go along, you know?

jbenet · 2014-11-17T08:27:52Z

Dockerfile


 EXPOSE 4001

-ENTRYPOINT ["ipfs"]
+ENTRYPOINT ["ipfs2"]


let's make a note of changing this back to cmd/ipfs and add it to the milestone once this PR is done.

This commit updates go-datastore in order to access new LevelDb Close() method License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@mappum

@mappum @jbenet LGTU? License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

allow framework to add level to log so level can be applied when using text formatter License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@jbenet

@jbenet @whyrusleeping @mappum If we permit initialization in `ipfs daemon`, then we must ensure that the node instantiated in `ipfs init` (to create the welcome file) shuts down and releases resources. Kept running into "resource temporarily unavailable". Discovered that it was our cousins Jeff and Sanjay over at LevelDB Ave. go-datastore doesn't expose Close() so I extended the TsDs interface and submitted a patch. License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com> Signed-off-by: Brian Tiger Chow <brian.holderchow@gmail.com>

jbenet · 2014-11-17T08:47:03Z

cmd/ipfs2/ipfs.go

+ initCmd: cmdDetails{doesNotUseConfigAsInput: true, cannotRunOnDaemon: true, doesNotUseRepo: true},
+
+ // daemonCmd allows user to initialize the config. Thus, it may be called
+ // without using the config as input


+1 for listing here the reasoning behind each option. it's important to track that somewhere. we got sloppy in the rush

jbenet · 2014-11-17T09:05:33Z

@maybebtc cool PR! Some comments above.

Here's a checklist to make sure to address:

btc · 2014-11-17T09:31:41Z

I'd like to propose that we make a conscious decision to use our errors package as errors. The errors are helpful enough that the practice may pay off in a big way (in terms of ease of development) if we can get religious about using them (internally) instead of the standard error.

~~Previously, they were debugerror, but this was a bit long.~~

reverting to debugerror. makes sense.

jbenet · 2014-11-17T09:51:55Z

@maybebtc i can get behind using our own errors package. Just note there is a cost in naming it the same thing: it confuses outsiders. Agree that debugerror is too long. maybe ierr, ierrors, iperrs, dbgerr? (and it doesnt matter what you name the directory itself. the beauty of explicit imports :) )

btc · 2014-11-17T10:05:41Z

I'd prefer to stick to generic names for the generic, generally-applicable packages. Keeps us honest as we try to break apart these generic libraries and make them available for use outside of the project. With that in mind, I'm open to any name we choose.

btc · 2014-11-18T05:18:12Z

ping @jbenet

jbenet · 2014-11-18T05:58:25Z

@maybebtc check it out: 6b2adf8!

jbenet · 2014-11-18T06:03:01Z

@maybebtc LGTM too. sec, i'm figuring out the error

jbenet · 2014-11-18T06:12:43Z

Interesting, the test really broke. ipfs cat without mount is broken-- it only returns 256K of data. wat.

jbenet · 2014-11-18T06:18:46Z

Well, fun:

> random 104857600 42 >bigfile
> ipfs2 add bigfile
added QmWXysX1oysyjTqd5xGM2T1maBaVXnk5svQv4GKo5PsGPo bigfile
> ipfs2 cat QmWXysX1oysyjTqd5xGM2T1maBaVXnk5svQv4GKo5PsGPo > bigfile2
> diff bigfile bigfile2
Binary files bigfile and bigfile2 differ
> du *
100M    bigfile
256K    bigfile2

@jbenet

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com> # TYPES # feat # fix # docs # style (formatting, missing semi colons, etc; no code change): # refactor # test (adding missing tests, refactoring tests; no production code change) # chore (updating grunt tasks etc; no production code change) Signed-off-by: Brian Tiger Chow <brian.holderchow@gmail.com>

will support daemon initialization License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com> # TYPES # feat # fix # docs # style (formatting, missing semi colons, etc; no code change): # refactor # test (adding missing tests, refactoring tests; no production code change) # chore (updating grunt tasks etc; no production code change) Signed-off-by: Brian Tiger Chow <brian.holderchow@gmail.com>

@jbenet

@jbenet @whyrusleeping @mappum very helpful for tracking down errors. the stack traces are only shown when debug mode is visible. They function best when caught at the source. I propose we use this errors package as a drop-in replacement for fmt.Errorf and errors.New in all of our code, and use errors.Wrap for external errors as they emerge from others' libraries. License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com> Signed-off-by: Brian Tiger Chow <brian.holderchow@gmail.com>

@jbenet

@jbenet now, ipfs can be built and executed in one step: ``` docker run jbenet/go-ipfs daemon --initialized ``` License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@jbenet

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@jbenet

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@jbenet

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

@jbenet

@jbenet yeah, these are just sane defaults License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

this is silly but whatever License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

looking forward to fixing these again License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

Warning: during normal execution node teardown must be the last thing that happens because command requests return io.Readers, which may still be constructing or processing their output. The node (and its subservices) is needed for this. good night and good luck.

feat(daemon) `ipfs daemon --initialized`

btc added the status/in-progress In progress label Nov 16, 2014

btc force-pushed the feat/daemon-init branch 2 times, most recently from af3be64 to 68613ee Compare November 16, 2014 21:51

jbenet reviewed Nov 17, 2014
View reviewed changes

Brian Tiger Chow added 7 commits November 17, 2014 00:43

deps(go-datastore) update

5fa31f8

This commit updates go-datastore in order to access new LevelDb Close() method License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

refactor(errors) debugerror -> errors

cf8fb87

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

feat(util) FileExists

11fd2aa

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

fix(daemon) if lock file does not exist, then the lock is not acquired

356a2d7

@mappum @jbenet LGTU? License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

deps(eventlog) no need to augment Logger

884d629

allow framework to add level to log so level can be applied when using text formatter License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

refactor(eventlog) use polite formatter as a direct dependency

ddb0189

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

jbenet reviewed Nov 17, 2014
View reviewed changes

btc force-pushed the feat/daemon-init branch from 68613ee to 7dd3813 Compare November 17, 2014 10:08

Brian Tiger Chow and others added 20 commits November 17, 2014 22:36

feat(util) add datastore Closer Wrapper

bacf3ec

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

feat(eventlog) initialize event logger with silent defaults

aba4b2b

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

feat(2/daemon) --initialized

0612004

@jbenet now, ipfs can be built and executed in one step: ``` docker run jbenet/go-ipfs daemon --initialized ``` License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

docs(eventlog/formatter)

94bfc81

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

fix(2/daemon) initialized --> init

bc39661

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

revert to debug error

ef65bb1

@jbenet License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

doc(eventlog) init

84354c7

@jbenet yeah, these are just sane defaults License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

docs(Dockerfile) update docker 'run' example

cc02471

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

delegate teardown

24f0fa0

this is silly but whatever License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

inline prints

7cebc3c

looking forward to fixing these again License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

fix

8d4c216

License: MIT Signed-off-by: Brian Tiger Chow <brian@perfmode.com>

test: test init output

493e5d3

test: cat test actual output

088016c

fix node context (tests pass)

d93e49e

jbenet force-pushed the feat/daemon-init branch from 1db6e5d to eba0599 Compare November 18, 2014 07:16

jbenet added a commit that referenced this pull request Nov 18, 2014

Merge pull request #355 from jbenet/feat/daemon-init

6107612

feat(daemon) `ipfs daemon --initialized`

jbenet merged commit 6107612 into master Nov 18, 2014

jbenet removed the status/in-progress In progress label Nov 18, 2014

jbenet deleted the feat/daemon-init branch November 18, 2014 07:22

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(daemon) `ipfs daemon --initialized` #355

feat(daemon) `ipfs daemon --initialized` #355

btc commented Nov 16, 2014

whyrusleeping commented Nov 17, 2014

whyrusleeping commented Nov 17, 2014

btc commented Nov 17, 2014

jbenet Nov 17, 2014

jbenet Nov 17, 2014

jbenet commented Nov 17, 2014

btc commented Nov 17, 2014

jbenet commented Nov 17, 2014

btc commented Nov 17, 2014

btc commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

feat(daemon) ipfs daemon --initialized #355

feat(daemon) ipfs daemon --initialized #355

Conversation

btc commented Nov 16, 2014

whyrusleeping commented Nov 17, 2014

whyrusleeping commented Nov 17, 2014

btc commented Nov 17, 2014

jbenet Nov 17, 2014

Choose a reason for hiding this comment

jbenet Nov 17, 2014

Choose a reason for hiding this comment

jbenet commented Nov 17, 2014

btc commented Nov 17, 2014

jbenet commented Nov 17, 2014

btc commented Nov 17, 2014

btc commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

jbenet commented Nov 18, 2014

feat(daemon) `ipfs daemon --initialized` #355

feat(daemon) `ipfs daemon --initialized` #355