Replication tips to speed up replication

[XIANLEE78]

Please forgive me if this was already answered but I could not find it.
I have read through the FAQ, product documentation, and searched the web and forums but did not find what I was looking for.
Tried searching for offsite+replication+best but that just took me in circles with others looking for that same "document"

Getting ready for replication. I have a good handle on how to configure and test but I have a couple questions.

Does it matter and if so which is better for deduplication and replication, thin or thick provisioned vmdk

thin replication to thin or thick replication to thick, which is better for replication when it comes to deduplication

replicating over wan so I am trying to determine which, if there is any difference, would make for the smallest replication of daily changes

I will seed replica from backup on external for initial seeding of destination so only daily changes need to be replicated, does disk provisioning matter for this

Any advice is appreciated, Thanks!

[dellock6]

I think replicate to thin gives you savings on the target storage, but since basically every activity is managed via incremental backups (both via CBT or via hashing the VM disks) it does not matter which one you choose for replicating, only modified data will be moved.

Luca.

[XIANLEE78]

@Luca

Thanks. Thats kind of what I thought but I wasn't sure if veeam was able to "process/compare" one type faster/easier than the other or not.

Backup > Seed Replica > first run after succeed seed > "Processes entire max size of thin disk" during the replication to determine modified data, so even though only 100 of the 500 GB provisioned is used (actual used space on datastore) it appears to "processes" all 500GB when determining motified data as veeam sees "provisioned" sizes so it reports 500GB XGB at whatever speed so X% complete

sorry if that makes no sense, I'm just confused why the "comparison" part takes so long

[Gostev]

Processed counter will always use full VM size for reporting purposes, but it will jump through zero block areas fast - if you watch closely, you will see it jumping when it hits those areas of the virtual disk.

Refer to read and transferred counters for the actual amount of changed blocks as reported by VMware, and actual amount of data transferred over WAN (after dedupe and compression).