Skip to main content

ETCD: POST vs. PUT understanding

ETCD is distributed key value store used as a core component in CoreOS. I've already send a post earlier this week. Here is a page describing how to use ETCD basic commands = ETCD API. Code snippets placed in a page mostly use put, but ETCD allows to use post as well. Most of us understand differences between those two commands in a notion of a REST(ful) service, but how does it work in key value store?

POST

Example over many words.

curl -v http://127.0.0.1:2379/v2/keys/test -XPOST -D value="some value"
curl -v http://127.0.0.1:2379/v2/keys/test -XPOST -D value="some value"

Two same command result into following content:

1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
{
  "action": "get",
  "node": {
    "key": "/test",
    "dir": true,
    "nodes": [
      {
        "key": "/test/194",
        "value": "",
        "modifiedIndex": 194,
        "createdIndex": 194
      },
      {
        "key": "/test/195",
        "value": "",
        "modifiedIndex": 195,
        "createdIndex": 195
      }
    ],
    "modifiedIndex": 194,
    "createdIndex": 194
  }
}

So ETCD adds an index value and put it into resulting key - which is also path to the value. For instance:

curl -v http://127.0.0.1:2379/v2/keys/test/194 -XGET

Allows you to get the specific key. The index is explicitly expressed in the url.

PUT

Put command just add or update given key. Let say I would use following example:

curl -v http://127.0.0.1:2379/v2/keys/test -XPUT -D value="some value"

Resulting content on test key is expected.

1
2
3
4
5
6
7
8
9
{
  "action": "get",
  "node": {
    "key": "/test",
    "value": "",
    "modifiedIndex": 198,
    "createdIndex": 198
  }
}

How to Model Add and Update Method?

My current task is to model and implement repository using ETCD under the hood. Usual repository contains CRUD methods for particular set of entities. Reasonable approach is to separate add from update to do not replace existing object, e.g. when using optimistic locking.

I don't want to see revision - index - numbers within keys so post command is not useful here. ETCD brings prevExist parameter for this use cases.

I want to perform add method which expect that there is no content on given key. I'll use following statement:

curl -v http://127.0.0.1:2379/v2/keys/test?prevExist=false -XPUT -D value="some value"

When you did not delete the key, as I did not, you can get following error:

1
2
3
4
5
6
{
  "errorCode": 105,
  "message": "Key already exists",
  "cause": "/test",
  "index": 198
}

On the other hand, use false to express update existing entity.

curl -v http://127.0.0.1:2379/v2/keys/test?prevExist=true -XPUT -D value="some value"

This command results into positive response.

< HTTP/1.1 200 OK

The repository uses put for both add and update methods but value for prevExist is the difference.

Comments

asnvuags006 said…
Dear.I am very honored to introduce a free date/photo recovery products to you, it is very useful and
free,and professional technical support.THANK YOU !

Popular posts from this blog

NHibernate performance issues #3: slow inserts (stateless session)

The whole series of NHibernate performance issues isn't about simple use-cases. If you develop small app, such as simple website, you don't need to care about performance. But if you design and develop huge application and once you have decided to use NHibernate you'll solve various sort of issue. For today the use-case is obvious: how to insert many entities into the database as fast as possible?

Why I'm taking about previous stuff? The are a lot of articles how the original NHibernate's purpose isn't to support batch operations, like inserts. Once you have decided to NHibernate, you have to solve this issue.

Slow insertion
The basic way how to insert mapped entity into database is:
SessionFactory.GetCurrentSession().Save(object);But what happen when I try to insert many entities? Lets say, I want to persist
1000 librarieseach library has 100 books = 100k of bookseach book has 5 rentals - there are 500k of rentals It's really slow! The insertion took exactly

Tomcat 7 remote deployment

I decided to provide automatic deployment of war packaged application using Jenkins and Deployment plugin. The target platform is Amazon with Tomcat 7, see nice set of articles to find out how to setup such environment for free.

Well, there is couple of tutorials but they missing some points so it pushed me to lost one hour of my work.

What I gotFresh installation of Tomcat 7 on remote machine with opened 8080 port on firewallPersonal war file supposed to be deployed
How to push it to tomcat?
1. First of all, there is simple configuration of tomcat users in file tomcat-users.xml - it was my pain in the ass :-) As original comprehensive documentation says, it's necessary to define user, but which one(s)?

Here is working example of tomcat-users.xml:

<tomcat-users>
<user username="manager-gui" password="changeit" roles="manager-gui"/>
<user username="manager-script" password="changeit" roles="manager-script"/>

Git on Windows: MSysGit

I have started to use Git today. I read a lot of discussions that there is no good tool for Windows platform. After forethought I have decided to used TortoiseGit. I also feared of difficult work related with Git as a lot of articles mentioned many instructions. As I already said, I have decided to use TortoiseGit, because I'm used to work with TortoiseSvn, but for start, MSysGit is enought. So this article is about MSysGit, next will be about TortoiseGit.

How to start with MSysgit on local machine?
Download and install Git for WindowsCreate source code directory for your git appRight click the directory at your favorite file browser. Menu should contain item "Git init here". It initializes chosen directory to be git-abled :-)It was your first usage of Git.

Commit data to local Git repository

Now, you can add any file, your first source code, to created directory. If you are prepared to commit any changes to your local git repository, follow next instructions.
Right-click th…