Low level etcd v2 client written in Javascript with failover support
npm install @npmcorp/etcdjsnpm fork of etcdjs
Low level etcd v2 & v3 client written in Javascript with failover support
``
npm install @npmcorp/etcdjs
Pass a connection string
` js
var etcdjs = require('@npmcorp/etcdjs')
var store = etcdjs('127.0.0.1:4001')
store.set('hello', 'world', function(err, result) {
store.get('hello', function(err, result) {
console.log('hello:', result.node.value)
})
})
`
If you have more than run instance of etcd running you can pass an array to load balance
` js`
var store = etcdjs(['127.0.0.1:4001', '127.0.0.1:4002', '127.0.0.1:4003'])
If you have a discovery token from https://discovery.etcd.io/ you can also pass that
` js`
var store = etcdjs('https://discovery.etcd.io/my-token')
etcdjs will automatically refresh its internal host list every 30s so you can transparently
add more machines to your cluster without updating your seed host list.
#### store = etcd(host, opts)
host should be a etcd host (or an array of hosts) and opts default to
` js`
{
refresh: false, // refresh the interval host list automatically
timeout: 60 * 1000, // default timeout for ops
json: false // stringify/parse all values as JSON
}
Note: The refresh option will try to discover additional etcd hosts via the etcd /machines endpoint which may not always return hostnames which are routable. Make sure the endpoint returns what you expect before turning on this feature.
#### store.get(key, [opts], cb)
Get a key. opts defaults to
` js`
{
recursive: false,
sorted: false,
json: false, // parse value as JSON for this request
wait: false,
waitIndex: (none)
}
#### store.set(key, value, [opts], [cb])
Set a key. opts defaults to
` js`
{
ttl: (none),
dir: false,
json: false, // stringify value as JSON for this request
prevExist: (none),
prevValue: (none),
prevIndex: (none)
}
#### store.update(key, value, [opts], [cb])
Set a key if it already exists. Same as set(key, value, {prevExists: true})
#### store.del(key, [opts], [cb])
Delete a key. opts defaults to
` js`
{
recursive: false,
dir: false,
prevExist: (none),
prevValue: (none),
prevIndex: (none)
}
#### store.mkdir(key, [opts], [cb])
Create a directory. Same as set(key, null, {dir: true})
#### store.rmdir(key, [opts], [cb])
Remove a directory. Same as del(key, {dir: true})
#### store.wait(key, [opts], [cb])
Wait a key to change. Same as get(key, {wait: true}) except the callback is called with a third argument next that will wait for the next change.
` js`
store.wait('hello', function onchange (err, result, next) {
console.log('change!', result);
next(onchange); // next will set waitIndex so we do not miss events
});
.wait returns a destroy function which can be used to kill a waiting request.
` js
var destroy = store.wait('hello', function onchange (err, result, next) {
// ... do stuff ..
})
destroy()
store.set('key', 'value') // won't trigger the wait
`
#### store.compareAndSwap(key, value, prevValue, [opts], [cb])
Only set if prevValue matches previous value. Similar to set(key, value, {prevValue: prevValue})
#### store.compareAndDelete(key, prevValue, [opts], [cb])
Only delete if prevValue matches previous value. Similar to del(key, value, {prevValue: prevValue})
#### store.push(key, value, [opts], [cb])
Create an in-order key that is guaranteed to be greater than the previous push. Check result.key to see the actual key.
#### store.machines(cb)
Returns an array of all machines in the cluster
#### store.leader(cb)
Returns the leader of the cluster
#### store.destroy()
Destroy the client and all open connections
#### store.stats.self([node], cb)
Returns node stats
#### store.stats.store(cb)
Returns store stats
#### store.stats.leader(cb)`
Returns leader stats
MIT