WARNING: Work in progress! Does not work yet!
While this port will probably also work on Solaris (most of it anyway ;-) it is untested.
This build was done in a base-64
19.3.0 (b81bcc1e-f1c8-11e9-8876-03a841507d8c
) zone which is based on pkgsrc 2019Q3.
Install prequisites:
# pkgin install gmake gcc7 git-base cscope yasm go libtool automake cmake bison flex curl unzip check pkgconf nss lz4 lzo doxygen
Checkout source:
# git clone --recurse-submodules https://github.com/siepkes/edgefs-smartos.git
# cd edgefs-smartos
# git checkout illumos-support
I needed to run the following command as a workaround for autoconf problems with pkgconf I ran into with libqb
(See: https://stackoverflow.com/questions/6758442/pkg-check-modules-breaking-in-solaris ):
# aclocal
Building:
# export NEDGE_HOME=/root/edgefs-dist
# export CFLAGS="-lsocket -lxnet"
# mkdir -p $NEDGE_HOME
# make deps
EdgeFS is high-performance, low-latency, small memory footprint, decentralized data fabric system released under Apache License v2.0 developed in C/Go.
EdgeFS is built around decentralized immutable metadata consistency to sustain network partitioning up to many days and yet provide consistent global data fabric namespace via unique versions reconciliation technique.
EdgeFS advanced globally immutable versioning with file-level granularity unlimited snapshots, global (LAN/WAN) data deduplication and geo-transparent data access enables variety of decentralized Edge/Fog/Cloud Computing use cases.
It integrates with Kubernetes and Docker Compose. Enables Edge/Fog/Cloud applications to use Amazon S3 compatible API with AI/ML S3X enhancements, built-in NoSQL Database, scale-out File via NFS, scale-out block via iSCSI/NBD.
EdgeFS is capable of spanning unlimited number of geographically distributed sites (Geo-site), connected with each other as one global name space data fabric running on top of Kubernetes platform, providing persistent, fault-tolerant and high-performance fully compatible S3 Object API and CSI volumes for stateful Kubernetes Applications.
At each Edge location, EdgeFS segment nodes deployed as containers (Kubernetes StatefulSet or Docker Compose) on physical or virtual nodes, pooling available storage capacity and presenting it via compatible S3/NFS/iSCSI/etc storage emulated protocols for cloud-native applications running on the same or dedicated servers.
EdgeFS can run on top of any underlying storage architecture - high-performance NVMe Fabric, local raw devices, existing filesytems or high-performance key-value databases.
Essentially it is high-performance versioning system that is capable of achieving 1,000,000+ versions per second rates, where each version is universally identifieable and location independent.
If you familiar with "git", where all modifications are fully versioned and globally immutable, it is highly likely you already know how it works at its core. Think of it as a world-scale copy-on-write technique. Now, if we can make a parallel for you to understand it better - what EdgeFS does, it expands "git" paradigm to object storage and making Kubernetes Persistent Volumes accessible via emulated storage standard protocols e.g. File, Object, NoSQL databases and even block devices, in a high-performance and low-latency ways. With fully versioned modifications, fully immutable metadata and data, users data can be transparently replicated, distributed and dynamically pre-fetched across many Geo-sites.
- Development Guide - details on build environment
- Harness Tests - details on how to run Harness Test package
- Minimalistic CLI tool - CLI tool implemented in Go
- LIBCCOW API - generic client library and its direct access API
- NFS and FSIO API - POSIX file system layer API
- Block and LIBCCOWVOL API - Block device API and iSCSI/NBD integration
- Edge-X S3 API - details on advanced S3-compabile features, e.g. Versioned Edits, Snapshots, etc
- FIO Engines - details on how to run FIO benchmarking tool at various directly conntected I/O layers
- LIBPMU userspace networking - details on libpmu, DPDK-like library
Make sure Docker package is installed on your build server with version >= 17.05 with support for staged builds. Execute the following command:
git clone git@github.com:Nexenta/edgefs.git cd edgefs docker build -t edgefs .
Deployments:
Configurations:
- Initialization - generic initialization procedure, applicable
- Kubernetes RookIO Deployment - segment deployment procedure
- Kubernetes RookIO CSI Integration - detailed instructions on how to get CSI configured with EdgeFS RookIO
- Kubernetes RookIO Monitoring - Prometheus and Graphana integrations
- Kubernetes RookIO VDEV Management - disk/VDEV health checking, replacement, addition, etc
- Kubernetes RookIO Upgrade - detailed instructions on how to execute rolling upgrade
- Kubernetes RookIO GUI - segment dashboard, CRD wizard
- Kubernetes RookIO NFS - setting up Scale-Out NFS (File)
- Kubernetes RookIO S3 - setting up AWS S3 compatible interface (Object)
- Kubernetes RookIO S3X AI/ML - setting up S3X interface for AI/ML, NoSQL and other intensive low latency workloads
- Kubernetes RookIO iSCSI - setting up Scale-Out iSCSI (Block)
- Kubernetes RookIO ISGW (Global Namespaces) - setting up geo-transparent capable global name space
- Kubernetes RookIO OpenStack/SWIFT) - setting up OpenStack/SWIFT inteface (Object)
- Users group: edgefs-users
- Slack channel: slack channel
- Twitter: @edgefs_io
- Telegram: edgefs
A regular community meeting takes place monthly on every first Wednesday at 9:00 AM PT (Pacific Time). Convert to your local timezone.
Any changes to the meeting schedule will be added to the agenda doc and posted to Slack #users and the edgefs-users mailing list.
Anyone who wants to discuss the direction of the project, design and implementation reviews, or general questions with the broader community is welcome and encouraged to join.
- Meeting link: https://zoom.us/j/404796463
- Current agenda and past meeting notes
- Past meeting recordings - coming