Hello all.
I'm relatively new to FreeNAS, but from what I'm seeing so far, it looks like a fantastic product and could be a great tool for a project I'm working on.
Here's what I'm looking to do. I'm wanting to have several FreeNAS boxes setup in multiple locations throughout the US (eventually the world) where I can take the information and keep it synced with each other. Now, I'm not set on any particular methodology yet, so here's what I've considered so far. Initially, we are looking to have 3-4 FreeNAS boxes that we'd want to keep synchronized, but this number could grow to be dozens over the next couple of years. This factor could almost immediately point me towards the first option, unless there is another option out there that I can look at. As far as the 'sync' method, I'm open to either 'Remote Replication' or 'Rsync' as both have their pluses and minuses.
Spoke-and-hub methodology - We setup a 'central' FreeeNAS that all of the systems sync through which then performs scheduled pushes and pulls to get the data fully updated.
Advantage:
- Smaller downtime between syncs on remote servers
Disadvantage:
- Central server going down could break the sync with the remote servers. This could be over-ridden either with a cluster at the hub and/or redundant connections setup between the servers to allow for sync if the central location is down.
Ring methodology - We setup Server A <-> B <-> C <-> D <-> A for the initial four boxes and as more get added, we insert them in between D and A, then E and A, etc.
Advantage: No dependency on a central server/cluster (helps cut down on costs)
Disadvantage:
- One server going down could mean that data from two servers that were only one hop away could then grow to a significant amount of time to replicate their information back and forth.
- When we get to the point of having a dozen plus systems, replication to get one piece of information between two servers on opposite ends of the 'circuit' could take a long time.
Anyone have any recommendations for how I could look at this? If I'm being an idiot and missing something very obvious, please let me know. Here's the answers to 'obvious' questions that I've had so far:
Q - Why wouldn't we want a solution from a vendor that can do this already?
A - We want a solution that we would have 100% control over, allowing for our own customization and security that could be changed at the drop of a hat.
Q - Why FreeNAS?
A - They are just the first one I've started with. I will be checking out other open source solutions to see what option will be the best for us.
Q - Why open source?
A - Why else? Access to the source code, the ability to make whatever modifications we'd want to make, the fantastic support community out there.
Thanks in advance to anyone that can help direct me on this project.
Charles
I'm relatively new to FreeNAS, but from what I'm seeing so far, it looks like a fantastic product and could be a great tool for a project I'm working on.
Here's what I'm looking to do. I'm wanting to have several FreeNAS boxes setup in multiple locations throughout the US (eventually the world) where I can take the information and keep it synced with each other. Now, I'm not set on any particular methodology yet, so here's what I've considered so far. Initially, we are looking to have 3-4 FreeNAS boxes that we'd want to keep synchronized, but this number could grow to be dozens over the next couple of years. This factor could almost immediately point me towards the first option, unless there is another option out there that I can look at. As far as the 'sync' method, I'm open to either 'Remote Replication' or 'Rsync' as both have their pluses and minuses.
Spoke-and-hub methodology - We setup a 'central' FreeeNAS that all of the systems sync through which then performs scheduled pushes and pulls to get the data fully updated.
Advantage:
- Smaller downtime between syncs on remote servers
Disadvantage:
- Central server going down could break the sync with the remote servers. This could be over-ridden either with a cluster at the hub and/or redundant connections setup between the servers to allow for sync if the central location is down.
Ring methodology - We setup Server A <-> B <-> C <-> D <-> A for the initial four boxes and as more get added, we insert them in between D and A, then E and A, etc.
Advantage: No dependency on a central server/cluster (helps cut down on costs)
Disadvantage:
- One server going down could mean that data from two servers that were only one hop away could then grow to a significant amount of time to replicate their information back and forth.
- When we get to the point of having a dozen plus systems, replication to get one piece of information between two servers on opposite ends of the 'circuit' could take a long time.
Anyone have any recommendations for how I could look at this? If I'm being an idiot and missing something very obvious, please let me know. Here's the answers to 'obvious' questions that I've had so far:
Q - Why wouldn't we want a solution from a vendor that can do this already?
A - We want a solution that we would have 100% control over, allowing for our own customization and security that could be changed at the drop of a hat.
Q - Why FreeNAS?
A - They are just the first one I've started with. I will be checking out other open source solutions to see what option will be the best for us.
Q - Why open source?
A - Why else? Access to the source code, the ability to make whatever modifications we'd want to make, the fantastic support community out there.
Thanks in advance to anyone that can help direct me on this project.
Charles
