Skip to content

Latest commit

 

History

History
170 lines (134 loc) · 5.37 KB

README.md

File metadata and controls

170 lines (134 loc) · 5.37 KB

PyPi version

krest

The Kaminario REST (krest) is a client library that provides ORM like interface for working with Kaminario K2 REST API.

Krest is written in Python and is aimed to provide rapid enablement of managing and monitoring Kaminario K2 all-flash arrays using Python.

This readme file complements the Kaminario K2 REST API guide document. You can request this document from the Kaminario support team.

Installation

pip install krest

Usage

Below are the examples of using krest. Again, we'll mostly explain how the Python interface maps to a URL spec which is outlined in K2 REST API guide.

First you need to obtain an endpoint (connection) to your K2 machine:

import krest
ep = krest.EndPoint("IP address", "username", "password", ssl_validate=False)

If you configured your K2 machine with real SSL certificates, set ssl_validate=True in the above call.

Creating and changing objects

Once you've obtained an endpoint, you can use it to CREATE/READ/UPDATE/DELETE objects.

# create host-group
hg = ep.new("host_groups")
hg.name = "hg1"
hg.save()

Parameters can also be passed inline:

# save() also returns the updated object
host = ep.new("hosts", name="h1", type="Linux", host_group=hg).save()

Note how we can use our hg object as host-group reference in newly created volume above.

Changing is simple - just change attributes and hit .save()

host.type = "Windows"
host.name = "h2"
host.save()

Searching stuff

If you know a specific object id, you can .get() it:

vol = ep.get("volumes", 1)

Otherwise use .search() to retrieve multiple objects that match search query. In the simplest form:

rv = ep.search("hosts", name="h2") 

The returned object is a ResultSet object that has a .hits array containing RestObjects. The total number of matched objects is recorded in the ResultSet.total attribute.

.search() method recieves resource_type, options and query keyword arguments. Each query argument is treated as a field name and its value and a requested field value. i.e. in the above example, we search for hosts having name="h2".

You can add search modifiers to field names:

rv = ep.search("events", level="INFO", message__contains="h1", name__contains="HOST")

For the full list of search modifiers please refer to the K2 REST API guide.

Notes on field values

  1. If a field value is an instance of RestObject, it is converted to its reference url and .ref is added to the field name. This allows native usage of RestObject in your code, i.e. ep.search("hosts", host_group=hg)
  2. If a field value is instance of ResultSet, list or tuple, then list elements are converted to string by comma-joining and __in is added to the field name. This allows doing things like ep.search("hosts", host_group=[hg1, hg2])

Working with ResultSets

ResultSets are returned by the .search() method of KrestEndPoint. The number of results returned from the API is limited to 100. You can check the total number of results matching your search query by inspecting the .total attribute of the result set.

For queries matching a large number of objects, you can use __limit and __offset query parameters to fetch results by chunks.

NOTE: Its crucial to sort results to retrieve objects in predictable order (use the __sort and __sort_order query attributes). All objects have an id field, to it is a good candidate to be used as a sorting field.

ResultSet is iterable, i.e. for r in rv:... is similar to for r in rv.hits:.... It also supports len(rv) and truthy evaluation.

If you set the .autofetch attribute of a ResultSet object to 'true' before iterating it, it will automatically fetch the next chunk of objects when the current chunk is exhausted. Don't forget to apply the sorting (as in the above note).

Deleting objects

Once you have a RestObject at hand, simply call its .delete() method to delete it.

More examples for object creation and manipulation

NOTE: All sizes in our K2 REST API are in kilobytes, with performance data being the only exception - it returns results in bytes.

Create a volume-group

vg = ep.new("volume_groups", name="vg1", quota=100*2**20)
vg.capacity_policy = ep.search("vg_capacity_policies").hits[0]  # search ad-hoc
vg.save()

Create a volume in the above volume-group

vol = ep.new("volumes", name="v1", size=10*2**20, volume_group=vg).save()

Map a volume to a host-group

mapping = ep.new("mappings", volume=vol, host=hg)
mapping.save()

LUN editing is easy:

mapping.lun += 10
mapping.save()

Map a volume to a host

host2 = ep.new("hosts", name="standalone", type="Linux").save()
mapping = ep.new("mappings", volume=vol, host=host2).save()

Create a snapshot

snap = ep.new("snapshots")
snap.source = vg
snap.retention_policy = ep.search("retention_policies").hits[0]
snap.short_name = "s1"   # Note - use short_name, and not just name
snap.save()

Create a replica from the snapshot and map it

rep = ep.new("snapshots")
rep.source = snap
rep.short_name = "r1"
rep.retention_policy = ep.search("retention_policies").hits[0]
rep.is_exposable = True
rep.save()
mapping = ep.new("mappings", volume=rep, host=hg).save()

Restoring a volume-group from a snap is a breeze:

vg.last_restored_from = snap
vg.save()