How to use EqualLogic's replication with DFS?

Question

Hi there,
We have file servers at 2 sites - HQ and manufacturing plant. They house users' home directories and other shared folders, and they are synchornized by Windows 2003 R2's DFS.
We will move the storage of these file servers to Equallogic - PS5000 at HQ and PS4000 at the plant.

How can we use Equallogic's replication features with Windows' DFS? I think's Equallogic's replication should be more bandwidth friendly than DFS, but I don't think I can totally abandon DFS, right?

Thanks in advance,
Chu

Donald_Williams · Answer

Hello,

To add a little but to what Joe said, one other facet of Replication is that the remote site volumes are not directly accessible.. So DFS couldn't see that as a volume with the same data as the primary side. The array is working at the block level,not file level as DFS does.

There is a product that makes DFS work really well and that's Polyserve with the DFS plugin. It would allow you to present the same EQL LUN to multiple servers at the same time. So DFS is always "in sync" it never has to send files back and forth. Then you can replicate those volumes to the DR site and mirror the setup there. Primary site goes down, turn up all the replicas and again with polyserve allowing concurrent access all the DFS nodes are already in sync.

Regards,

-don

Joe S586 · Answer

Hi, I’m Joe from Dell EqualLogic. No, this is not possible. The DFS is a Namespace, formerly known as Distributed File System, which allows administrators to group shared folders located on different servers and present them to users as a virtual tree of folders known as a namespace. A namespace provides numerous benefits, including increased availability of data, load sharing, and simplified data migration.

The Dell EqualLogic replication is a replication technology provided by PS Series firmware, where you can copy volume data from one group to another, protecting the data from a variety of failures, ranging from the destruction of a volume to a complete site disaster, with no effect on data availability or performance. Essentially, this is a block by block copy of changes to one volume to another location, with no ability to “merge” individual volumes together to create one massive volume.

Regards,
Joe

Chutikorn · Answer

Thank you, Joe. That was very clear. And thanks Don, I'll look into Polyserve, although I'm afraid I'm not gonna have budget for it.

Donald_Williams · Answer

You are very welcome. Glad to be of assistance.

Re: Polyserve. Yeah, last I knew we're talking thousands of dollars. But it works slick as can be. I got a training class on it and it was smoking fast once set up. And w/o the overhead of DFS moving stuff around it made DFS work as you expect.

SANBOLIC does MelioFS and that's had good reviews, less setup overall but I'm not sure they have the level of integration that Polyserve does.

http://www.sanbolic.com/

Take care.

-don

sanbolic · Answer

Don, thank you for bringing up our solution! We have outstanding integration with Microsoft DFS and provide scaleout of CIFS and NFS out of windows based servers. We would be happy to host a livemeeting or provide your customer with free evaluation. In addition to fileserving, we provide integration with Active Directory, VSS back up, Dynamic Expansion of volumes, and centralized provisioning and management of the network shares. You can see more here http://www.sanbolic.com/web_serving.htm, and our products are available through Dell. Please feel free to contact us if we could be of any help!

khsieh · Answer

I have a similar setup to yours with an EqualLogic group at HQ and a DR site and file servers (among others) at both locations. I am using EqualLogic replication to protect volumes in case of disaster, and I have DoubleTake protecting some SQL data, but I am using DFS for my file servers and I suggest that you do it too.

DFS Replication is more efficient than both EqualLogic and DoubleTake replication. With DFS, only changes get sent, at a byte level. With EqualLogic, each page that gets dirtied (written to) gets replicated, no matter how much of the page was modified. I don't know exactly how large a page is in EqualLogic, but I have heard that it is in the order of MB, so replication isn't nearly as efficient. The volume replica also isn't online as others have described, which doesn't make it very useful for your purposes.

I have tried DoubleTake, but it isn't designed to allow write access to both the source and target of the replication, which means that it isn't good for your use case. DoubleTake is also not very efficient at handling large volumes, because anytime the replication gets out of sync (say you had to reboot the source server), it has to rescan and compare every file in the replication set. That used to take me a day. DFS doesn't have this problem because the DFS-R database keeps track of changed files and where the changes are located, so syncing changes is very efficient, even is a member has been offline for a while.

If you have DFS-R working right now, I suggest that you stick with it. You can move the files to the EqualLogic, but use DFS-R for the replication. To protect your files first use shadow copies from Windows as the primary protection, and then traditional backups and/or EqualLogic snapshots.

Chutikorn · Answer

Awesome, khsieh. Nothing beats getting the experience of someone who's been through it already. Thanks a bunch.

jim_milano · Answer

I realize that this is an old thread but I came looking for the same kind of solution. I am not as thrilled about DFS-R as some of the earlier posters. I find that our users frequently leave files open for long periods of time and open files are not replicated. This results in mis-matched versions being updated by different users connected to different targets. Ultimately I ended up leaving one target 'disabled' and forcing everyone to connect to a primary target while leaving DFS-R replicating. I've also had replication database corruption on multiple occasions and recovery was messy and time consuming. I would be happy to use DFS for it's virtual file structure and ability to 'manually' fail users over to an alternate target attached to a replica of the live LUN. We used Dell/EMC storage 10 years ago and their 'Replication Manager' program would trigger a LUN replication and then mount the replica on a different host. It seems to me that this would have allowed me to achieve at least a 'warm' DFS target as a failover target that is left 'disabled' until I need it.

jacksonbilly · Answer

Hey Jim...I'm looking into replication using DFS or EqualLogic's Replication. I'm trying to figure out which volumes would be best by using DFS or EQL or even VMware's SRM. I am about to power up an EQL at a remote site, which we are calling our DR site. So I'm needing to replicate volumes from the EQL I have running at our corporate office to the EQL at our DR site. I would greatly appreciate any advice you would be willing to provide on what volumes to replicate using DFS, EQL or SRM. We have file servers, a single exchange server 2010, domain controllers, oracle servers, an old SQL 7 server, and will eventually have a sharepoint server.

I was told today that vmware has a 'Auto SnapShot Manager' software that is capable of making replication of exchange an 'ok' procedure to do. So it would coordinate the transaction logs with the database.

Thank you in advance,

George

FluidFS

How to use EqualLogic's replication with DFS?

Was this post helpful?

Register Now! for Secure Secrets with Kubernetes & DevSecOps - and grab a lab!