Deploying Synchronous Replication with Dell Equallogic PS4100XV to PS6010XV

Synchronous Replication can increase the load to the network, especially between the members and the switches they are connected to.  A good rule of thumb is to use a dedicated stacking link that provides bandwidth at a minimum of 20 Gbps full-duplex, or if using ISL make sure you have at least 12 Gigbit throughput to start (4100 2x's 1GbE Plus the 6010's 1x's 10GbE = 12Gb).

Make sure you review this document that discusses deploying mixed 1GB and 10Gb SANs: http://en.community.dell.com/techcenter/storage/w/wiki/2640.deploying-mixed-1-gb-10-gb-ethernet-sans-using-dell-equallogic-storage-arrays-by-sis.aspx

Both the Group Admin Guide and the GUI online help has good information on SyncRep configuration and performance monitoring.

Also the Configuration guide has a lot of useful general informaion :

http://en.community.dell.com/techcenter/storage/w/wiki/2639.equallogic-configuration-guide.aspx

Synchronous replication also requires that both members be in the same Group but the member arrays are in different pools.

-joe

Hi Joe

Thanks for your your feedback and help here.

I have decided to use 2 units of PS4100XV for the sync rep replication. Is there any guidelines or if there would be some network bandwith that might be experience since it only uses 2 x 1Gb throughput. Would you recommend syn rep or just a typical asynchronous replication.

If I'm going to use asynchronous replication, what is the nearest time to configure every replucation schedule (ex every 5mins, 10mins, hour). The distance of each array is about 400 meters.

Thanks again for your help

The Group admin guide has the information to ensure you have successful replication that meet your business needs, such as determining the number of replica's you need to keep, time between replica's, etc.

I really cannot recommend one method over the other for your business disaster recovery needs. 

Generally speaking:

Traditional replication is a point in time process that copies data between two groups usually in two different geographical locations. Replication provides protection against a disaster such as an earthquake or floods, etc.

SyncRep is a real-time process that simultaneously writes volume data across two different pools within the same PS Series group. So you are maintaining two copies of the volume data in the same data center, or to two different buildings on the same campus connected with high speed links.

Also keep in mind that your data should still be backed up using traditional methods and observe the 3-2-1 rule. Replication is great for quick recovery and offers a way to recover from disasters without the need to restore from tape or disk copies.

As far as the ISL using just the two PS4100's that you mentioned, I would bump the ISL up to 4Gb's (2 members X's 2 1GbE's = 4Gb), or just use a 20Gb stacking cable.

Again, the network load depends on several factors, such as volume traffic, network bandwidth, etc.  The best way to is to monitor your Array's performance with SANHQ, and adjust schedules, etc. accordingly.

When considering the replication schedule (for traditional replication), I typically consult with the line of business that owns the data to see what is the longest amount of time that they can tolerate having to re-enter data, i.e., if a system goes down, and you restore a replication (or snapshot) from 30 minutes prior to the outage, they would have to reenter 30 minutes of data between the outage and the time stamp of the replication/snapshot restore.

Also consider using shapshots which will give you additional restore options, and these can be setup on a schedule with a higher frequency.

 -joe

Hi Joe 

I am looking for documentation for async replication between a PS6100 to PS4100 in different locations. Any help would be appreciated.

Thanks Mike

Mike,

The best source for information on setting up Group to Group replication is the Group Admin Guide (eqlsupport.dell.com) and the Online Help from within the group manager (the little "?" in the upper right).  Also, prior to setting up replication it is vitally important to ensure that each member in the two groups can ping each other’s eth interface.  Use the ping and traceroute commands to verify this.

Ping:
 
To ping out of each of the specific ETH port interfaces use the commands listed below (test all members in the group to the other group both ways):
 
ping "-I "
 
(that is a –I as in Capital letter “eye”, AND ensure you use the quotes after ping and the end of the command!).

Traceroute:
GrpName>support traceroute -s [ETH port source IP] [destinationIP]
 
(make sure you test each ETH interface for all members in each group both ways)

Additionally, you cannot use a NAT firewall for replication, EQL replication requires that all network ports on each side are directly reachable.  The only ports required to be open for replication is 3260 (iSCSI) and (optional) the management network ports (if you setup a separate management network).

Lastly, the configuration portal has additional information on switch setup that should be reviewed.

en.community.dell.com/.../3615.rapid-equallogic-configuration-portal-by-sis
en.community.dell.com/.../2632.storage-infrastructure-and-solutions-team-publications

And the configuration guide:

en.community.dell.com/.../2639.equallogic-configuration-guide.aspx

Replication Sizing Guide and Best Practices:
en.community.dell.com/.../2641.dell-equallogic-auto-replication-best-practices-and-sizing-guide-by-sis

-joe