Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

OpenSVC was just ported to the FreeBSD platform. The little blurb in that announcement intrigued me so I went to the OpenSVC website and found this:

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

Plus some UML-type graphics. Which is all neat, but I still don't understand: what does it do? Am I just being dense? What's the use case for this system?

share|improve this question

3 Answers 3

up vote 2 down vote accepted

I'll try to describe specific cases to explain OpenSVC usefulness.

Consider a sysadmin in a corp, he sets up services for clients/users. He has about 50 services in charge. He likes FreeBSD, so he tends to deploy his services on that base. He has a good grasp of how rsnapshot works, so he created scripts to automate backups and dutifully prepared scripts to help recover from a server crash, may be even a site black-out.

The sysadmin in the next cubicle is also in charge of about 50 other services. He will have also done its homework right, but with its own style. He may prefer Linux and rsync, his recovery scripts will be in a different location (may be in its desktop). His clients may require more availabity, so he had to choose a clustering stack.

Now scale to tens of admins and thousands of services. The datacenter is a patchwork of technologies : 3 to 4 different os, 2 different storage hardware with their own replication protocol (hitachi shadow copy, emc srdf, netapp snapmirror), 2 clustering stacks (hacmp, redhat cluster, suncluster, veritas cluster), a myriad of different scripts to automate actions on small perimeters.

And picture some frequent scenarii: o leak over a rack: 20 servers down, 50 services needing failover, 10 different admins with all their specific failover mecanisms o site blackout : same sketch, a ten-fold o the corp has outsourced the service monitoring: hard to trust the low-profile screener with the fine-tuned services start/stop action responsability o sysadmins turn-over : all the fine-tuning is not easy to pass-over to the newcomer.

OpenSVC can be seen as a free, easy-to-deploy, deploy-everywhere cluster stack. Low criticity services can have only one node. Medium criticity services can have 2 nodes and no automatic failover. High criticity services, 2+ nodes with automatic failover plus a remote node for disaster recovery.

Same tool for all, respecting every sysadmins preference (os, virtualization model, filesystem, replication scheme) and every availablity target to provide the stop/start/replicate actions for such different types of integration.

I focused the example on a large scale environment to highlight OpenSVC usefulness, but in real life, many users use OpenSVC to manage 1 to 4 services, just to dish-out a lot of scripts they maintained themselves previously.

The web collector brings additional advantages, as a reporting, alerting and data-mining front-end. This component is not GPL'ed but it is not needed to benefit from the above. Free-lancers tend to use the internet collector to have a single point of reporting for services they maintain for differents clients.

Hope it help clarifying OpenSVC position in the clustering world.

share|improve this answer
    
Thanks for the great explanation –  sh-beta Jun 17 '10 at 17:37

Sounds like it's an aggregator for the status of services on a cluster of machines in a datacenter. Maybe like, a central place to monitor your file servers, web servers, NFS servers, virtual machines, etc. along with status logs and whatnot.

Additionally it sounds like you can restart the services, stop them, "ping" them, etc...basically a tool for helping control and monitor lots of computers in a datacenter from one place.

share|improve this answer

There are several cluster service managers for high availability Linux. (I don't have more links handy, though) This seems to be a FreeBSD centric offering to manage resources in a cluster setting (ie, make sure web server is available on at least 1 of the nodes in the cluster @ all times, etc)

share|improve this answer
    
Doesn't say much beyond the blurbs I already quoted and this is definitely not FreeBSD-centric - it was just ported a week ago. –  sh-beta Jun 17 '10 at 17:39

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.