This post is more than 5 years old
62 Posts
1
8286
Can someone explain how to interpret the meta data values from below output ?
Below is the output from couple of VMAX 250F arrays. Can someone explain me the difference between Array Usage Meta data % and Replication usage meta data %. What is the difference ? How can i calculate the total value ? In the second VMAX array usage Meta data is 100%. Does this mean this arrays needs to convert cache to meta data page space for future allocations ?
PedalHarder
465 Posts
0
January 24th, 2018 14:00
That is correct. A certain amount of metadata is required for devices and again it is a finite resource.
PedalHarder
465 Posts
1
January 24th, 2018 14:00
Replication cache is part of global memory that is used to manage local replication snapshots. This is a finite resource in that only so much capacity can be under local replication management. The metadata % provides the view to show the amount of replication cache that has been consumes. It is used to help you manage the resource.
Ayrton_Senna1
62 Posts
0
January 24th, 2018 14:00
Thanks you, what about Array usage Meta data % ?
I believe that is the portion of cache used for creating tdevs ? Once the Array Usage Meta data % reaches 100, i can no longer create new TDEV's.
Relication Meta Data : Portion of cache used for Vxsnaps. This part is clear.
Ayrton_Senna1
62 Posts
0
January 24th, 2018 14:00
Yes, i did some testing and it matches to what you said.
Now, it would be nice for the customer to see the breakdown of cache from symcli on these flash arrays. Cache seems to be bottle neck on flash arrays more than free space.
Whats the point of having Replication cache available when we cannot create target devices to link those snapshots ??
Something like Total System Cache MB -- This is visible in symcfg list
Array Usage Cache MB -- Only percentage is visible, it would be nice to know the net allocated value
Replication Cache MB -- Only percentage is visible, it would be nice to know the net allocated value
The percentages will certainly help us to monitor how much we can provision and how much we can setup local replication.