News

HA setup doc

June 28, 2010 23:11

A new documentation covering the high availability setup of an OpenSVC service using Open-HA has been added here

ZFS replication added

June 19, 2010 04:32

Starting with release 1.0-100619.1127 opensvc is able to drive zfs send/receive during syncnodes automation. This new data replication mode adds up to the already supported Drdb, Netapp snapmirror, EMC clones, snap+rsync and LVM2 delta replication modes.

AIX port merged

June 15, 2010 16:32

OpenSVC now supports hosted services on AIX nodes. IPv4 and IPv6 resources, jfs[2] and namefs mount resources, rsync based synchronization on jfs2 snapshots, and the usual EMC and NetApp replication modes. More to come with volume groups exports/imports across shared-disks cluster nodes, and WPAR containers. The rpm packages from the repository can be used for deployment.

AIX port coming

June 13, 2010 14:26

OpenSVC acquired an IBM Power server for the lab. The AIX port will start soon. WPAR support is targeted as a container technology. This port will open the doors of many datacenters to OpenSVC as an all-encompassing operations orchestrator.

FreeBSD packages now available

June 13, 2010 01:07

OpenSVC now delivers FreeBSD tbz packages at http://repo.opensvc.com/tbz/.

FreeBSD port merged

June 10, 2010 10:27

Starting with release 1.0-100609.1907 opensvc is able to drive hosted and jailed services on FreeBSD. Zfs is fully supported on local and shared disks. Zfs snapshots are created and removed as needed by opensvc for data replications between cluster nodes. Loopback devices are supported through mdconfig.

Drbd data replication driver merged

June 07, 2010 07:13

Starting with release 1.0-100607.0844 opensvc is able to drive primary/secondary role selection and connection establishment upon service startup or shutdown. This new data replication mode adds up to the already supported Netapp snapmirror, EMC clones, snap+rsync, zfs send/recv and LVM2 delta replication modes.

Sun LDOM virtualization driver merged

June 01, 2010 08:49

Starting with release 1.0-100601.0742 opensvc is able to start and stop services encapsulating a Sun LDOM guest. This new driver confirms opensvc ability to drive heterogenous datacenter operations : zones, kvm, hpvm, ldom. More to come.

Symmetrix clones support added

February 26, 2010 21:28

EMC Symmetrix arrays clones allow multi-target block device replication, splitting a set of cloned devices, resyncing and merging changes on a clone set to its source set. After the initial full copy, all resyncs are delta-based. These excellent features allow opensvc to drive services using a cloned device set, handling the splitting on start-up and providing a simple command for resync and merge-back. This kind of services are often used for maintenance, pre-production, performance testing environments, where the source device set is used for production.

Netapp snapmirror driver merged

February 09, 2010 22:49

Starting with release 1.0-100209.2116 opensvc is able to drive snapmirror operation on volumes allocated to a service. Such operation include spliting the sync or swaping the sync direction.

Brief

OpenSVC is a 'service' manager, as in clustered service manager, designed for real-world heterogeneous datacenters and large-scale operations orchestrator / automation (disaster recovery, for example).

Services are collections of resources (virtual machine, ip, disk groups, filesystems, file synchronizations, and application launchers).

Services can be started, stopped and queried for status, providing a consistent command set for wildly different service integration types.

Service configurations, status and logs are pushed to a central database coupled to a web front-end (collector).

Services can be administered using the stand-alone GPLv2 software stack deployed on the nodes (nodeware), or through the web-front end.

Feature Matrix

Nodeware

scalable: scales up to thousands of nodes and services, down to one node.
portable: runs today on Linux, FreeBSD, HP-UX, AIX, Solaris, Opensolaris and can easily be ported to any Unix-like operating system. Handles Windows service continuity through supported virtual machines.
versatile: can drive hardware-provided data replication for any kind of storage vendor (EMC SRDF, NetApp snapmirror). Can drive rsync, ZFS, Drdb based replications. Can drive service implemented as virtual machines (xen, kvm, lxc, zones, jail, Oracle Ldoms ...). Guest operating system may differ from the host's.
reliable: protects data through scsi3 persistent reservations, when active and supported on the hardware and operating systems. Disaster recovery plans can be described in service configurations. Optionally couples to standard heartbeat daemon (OpenHA, heartbeatd, keepalived) to provide higher service availability.
easy: a service container is created within minutes
open: the core components is licensed under GPLv2, the data collection and presentation website is a subscription-based component.
transparent: all system commands executed by OpenSVC are logged locally and centralized for audit
connected: bridges to site's asset manager and ticketing system for alarming and incident follow-up

Collector

The collector helps administrators to manage their services through a set rich feature set

dashboard: all site anomalies at a glance, one-stop page for morning checks
monitor: display status of service resources
service topology: sketches service-nodes-storage-sites relationships
actions: display coalesced and searchable action logs
disaster recovery plan: design disaster recovery plans action sequencing, audit auto-generated dr scripts
applications: assign group of responsibles to group of services, for alerts routing and privileges enforcement
nodes: display and edit asset information
alerts: display alert log (sent and queued)
nodes performance: graphs node metrics on user-select time ranges
service availability: graphs and compute per-service availability ranges and ratio
service control: run service and group of services actions directly from the collector
site statistics: show trends on the number of nodes, services, and overall quality of the site
absolescence management: set warning and alert thresholds on hardware and os and be notified of nodes obsolescence
packages incoherence alerts: raise alerts when members of the same cluster have different packages or package versions installed, which might keep applications installed as hosted services from finding infrastructure they need to work properly after a failover

The collector linked from this site's menu has its views filtered so that different users can share the same collector instance. As a consequence, you won't see services and nodes there unless we set up the appropriate privileges for your account. This service is not free, though trials can be requested at admin@opensvc.com. This shared collector is ideal for services hosted on off-site nodes (OVH, ...). For other cases, the prefered setup involves a private, on-site, collector instance.

OpenSVC datacenter model

typical opensvc cluster

opensvc service anatomy

opensvc node anatomy

opensvc site anatomy