Unsolved
This post is more than 5 years old
11 Posts
0
548
Seeding inital data of the RP
Hi,
According to the planing,we will do a initial seeding of data of the two Prd and DR Site RP at the local site for CRR replication.Once the finished of the RP replication data , then we will move the DR RP and VNX5200 array to the other remote Site.Is it feasible good solution and will need to chnage for the Remote IP of RP.eg.,IP address.Any concern and issue that need to take note/change according at the remote site RP.
forshr
1.1K Posts
0
April 27th, 2015 09:00
Hi there,
Set up your RP system for the seeding process and then prior to the move, pause the CGs. Before the hardware is moved to the DR site, use the Deployment Systems Modification Wizard to change the IP addressing and then once the hardware is in place and fully connected, resume the CGs. Note that prior to resuming the CGs I would recommend performing system connectivity checks via boxmgmt between the RPAs.
During the period of time that the CGs are paused, marking mode will be enforced, whereby new writes are marked by writing metadata to the Prod copy journal. When the CGs are resumed, the metadata is used to read the actual data from the Prod array. This can result in high read rate but can be throttled back as a safeguard against potential increased host latency. The period of time between pausing and resumption of the CGs can mean that a large amount of data is accumulated which can result in what looks like a full sweep, but it is not. It is a long resync without the comparative analysis of source and target LUN blocks that a full sweep performs. However, a long resync indicates that the amount of accumulated data is greater than the size of the replica copy journal and therefore the journal is bypassed with a single phase distribution of the data to the actual replica which is the default behaviour of a full sweep.
Regards.
Rich