Results 1 to 4 of 4
I spend a lot of time finding which cluster engine is best in my case(see requirements). Do you have any comments? Requirement is simple: 1. Linux only 2. is free ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
- 12-14-2012 #1
- Join Date
- Dec 2012
Cluster solution under linux
I spend a lot of time finding which cluster engine is best in my case(see requirements). Do you have any comments?
Requirement is simple:
1. Linux only
2. is free but has paid support
- 12-14-2012 #2
- Join Date
- Apr 2009
- I can be found either 40 miles west of Chicago, or in a galaxy far, far away.
- 12-15-2012 #3
- Join Date
- Dec 2005
CoroSync works very well as the Cluster Engine
In June of this year, I finished the R&D of a High-Availability storage system (For the sake of not incorrectly using the term "Cluster", which is actually three (3) or more servers. I built my HA system with two (2) servers after almost a year, which included selection and testing to find the most reliable solution.
I also started with a limited experience of Linux, as well as High-Availability storage systems.
Linux Distro: Ubuntu Server 11.10 (because I wanted Free as can be and 12.x was just coming out.)
HA/Cluster Layer: Pacemaker & CoroSync
Mirroring of disk block devices: DRBD (I'm using 8.3, due to availability in Ubuntu 11.10, and it is working very well)
A Java based management solution is also available.
These systems are used as NFS storage for a two head VMWare vSphere 5 infrastructure.
I recommend reading Linbit's Tech Guides and participating in the related forums and mailing lists.
When I setup my current systems, the tech guide on NFS storage:
linbit dot com > Support & Training > DRBD Support > Downloads > Tech Guides > highly-available-nfs-with-drbd-and-pacemaker
Was very helpful, but there were or are some important factors missing. That is where the mailing list came in handy.
I have since experienced a fail-over event, due to a power outage where one of the breakers tripped, it happened to the one the primary storage server is on. The fail-over was successful and bring the primary up as secondary also worked like a charm.
Let me know if you need any more info.
I'll leave you with this, "Fencing, fencing, fencing", I can't stress that enough.
- 12-15-2012 #4
are you looking for distributed file system kind of solution ? If so ,check out Glusterfs (now part of Redhat), Ceph (very promising) and Moosefs. (Heard its good one).All the above are free/open source and has paid support/contract.