Unsolved
This post is more than 5 years old
29 Posts
0
1967
RecoverPoint Consistency Groups
We use a vnx 5300 in our production environment as well as our DR environment with RecoverPoint in both places with a 20mb wan link connecting both.
In the RecoverPoint we have 3 consistency groups. Although all 3 groups occasionaly reach high load, their is one that is frequently at high load. What I'm trying to determine is why is there such a large amount of data being syncronized accross the WAN in such short time periods, seconds apart. If you look at the journal and sample images in literally seconds there are different images of MBs of data being replicated. Even large amounts 10mb, 19, 38, even 453 and 893mb! See screenshot. Where does the Sample image gather it's data from, especially the size in MB? I need to know what data in this group is actually being transferred and why so much. Thanks.
rupal1
433 Posts
0
August 7th, 2012 09:00
Transferring to the Recover Group for quick assistance on the same
Thanks
tolinrome1
29 Posts
0
August 8th, 2012 11:00
Anyone have an idea as to why such huge amounts of data increasing in size are being replicated within seconds? Thanks.
RickBrunner1
117 Posts
0
August 15th, 2012 16:00
Tolin,
The MB listed beside the image in the GUI does not necessarily represent the amount of data replicated, but the amount of data that would need to be modified on the replica if you selected that image, so it is the amount of changed data between one image and the previous image/time stamp. The reason that it does not represent replication is that you may have had dedupe and compression reduce this by upto 90x.
Also the GUI does not show all images that are in the journal, it is only a good representation across the complete timeframe of images available in the journal.
If you want to validate this simple select a point in time, select image access, then move forward or backwards to the next image. You will see that the next image is seconds to micro seconds apart depending on the change that occurred and the CG policy defined.
To understand the cause of the highload, run the detect_bottlenecks command interactively or via SSH shell.
ssh admin@x.x.x.x "detect_bottlenecks mode='Advanced and detailed general detection'"
You can specify a timeframe and the output will tell you how much data you replicated during the timeframe and if highload conditions existed, what is the bottleneck causing it ( wan, storage, RPA)