Don't want deduplication enabled on replica automatically

Software-based VM-centric and flash-friendly VM storage + free version

Moderators: anton (staff), art (staff), Max (staff), Anatoly (staff)

Post Reply
santiagocastro
Posts: 6
Joined: Thu May 07, 2015 8:52 am

Tue May 19, 2015 4:18 pm

Hi guys!

I started testing v8.0.792 a few days ago.

In my tests, I have noted when I create an high availability volume, deduplication on replica is enabled by default although it shows "Deduplication=no" on Starwind console.

I have edited .swdsk file of the volume on both nodes and I have seen that on first node it shows:

<DeduplicationEnabled>no</DeduplicationEnabled>
<SnapshotsDisabled>no</SnapshotsDisabled>
<BlockSize>0</BlockSize>

but on the replica the content is:

<DeduplicationEnabled>no</DeduplicationEnabled>
<SnapshotsDisabled>no</SnapshotsDisabled>
<BlockSize>4096</BlockSize>

When creating the volume, I used the following steps:

Right click on the first node/Add device (advanced)/HD Device/Virtual Disk/Specified location and size/This provisioned without Option deduplication checked/No L1 cache/No L2 cache/default targets.
Now right click on the new LSFS volume/Replication Manager/Add replica/Synchronous mode/Specified partner IP/CREATE A NEW PARTNER DEVICE/Specified partner device header and target name/No RAM disk/Specified network settings for the replica/Create replica.

Based on activity metrics checked using perfmon, I suspect when creating the volume using theese steps, Starwind always enables deduplication on the replica automatically. I observed when I copy the same file to the HA volume over and over, metrics shows write activity only on the first node after the first copy, and the memory consumption of the Starwind process is very high.

Please, did anyone see the same behaviour?
User avatar
Oles (staff)
Staff
Posts: 91
Joined: Fri Mar 20, 2015 10:58 am

Thu May 21, 2015 3:56 pm

Hello Santiagocastro!

May i ask you to provide me with next information:

• Are you sure you do not see any network traffic between nodes?
• What cache size are you using?
• How is this device connected and what operation system are you using?
• How is this device used and presented? (As Datastore,CSV,etc.)

Thank you.
Post Reply