Comprehensive data protection for all workloads
Post Reply
Argon
Influencer
Posts: 17
Liked: 1 time
Joined: Sep 29, 2015 1:17 pm
Contact:

Wan Accelerator troubles/questions

Post by Argon »

Hello,

I'm setting up the WAN accelerator on Veeam 10. I've read the Veeam KB's about it and i'm still confused (like many others it seems... :-) ).
To be specific I've configured the following as a test:
Target WAN: I've created a new local disk on that server: 500Gb in size and configured the Cache Size: 400Gb

Source WAN 1: Deployed the Source WAN accelerator and configured a disk with size: 30Gb and Cache Size: 30Gb
This Source has 8 VM's and and there is 1.25Tb in use

I've populated the Target WAN with a Windows Server 2016 OS. When I started the population, it immediately locked 400Gb on that Target WAN server. So there was still 100Gb free.

I started the Backup Copy Job between Source WAN 1 and Target Wan: all good !

Now I configured a second Source: Source WAN 2. I configured a disk of 120Gb in size and 100Gb in Cache size.
This locations has 11 VM's and is using 2.2Tb in total.

I started the Backup Copy Job between Source Wan 2 and Target Wan: On the target WAN there was another 95Gb added to the Global cache in a split second. So the Global Cache over there now is 495Gb while the disk is only 500Gb in size...

What am I doing wrong here?

The ultimate goal is to use the "Many to One" WAN accelerator situation since we need to replicate around 15 locations...
The 2 Source locations I took now are really small. We have some locations with 5Tb in size and around 40VM's. If I see how much data the Target Cache is already using for only the above 2 Source locations (which are rather small). The Global Cache on the Target side will be enormous...

Thanks!
foggy
Veeam Software
Posts: 21069
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Wan Accelerator troubles/questions

Post by foggy »

Hi, have you reviewed the WAN cache sizing recommendations in the user guide? It specifically mentions the many-to-one scenario.
Argon
Influencer
Posts: 17
Liked: 1 time
Joined: Sep 29, 2015 1:17 pm
Contact:

Re: Wan Accelerator troubles/questions

Post by Argon »

Hello,

Yes I did. But looks like i'm doing something wrong/don't understand what exactly to do...

Source 1 (low bandwidth mode): 5% of 1.25Tb = 62.5Gb (I've made a 30Gb disk for the Digest files as a test and it only used 10Gb right now. So i'm cheating here, but I can expand it whenever I want). Also the Source isn't the problem in my case, it's the target where I have troubles.

Source 2 (low bandwidth mode): 5% of 2.2Tb = 110Gb (I've made a 120Gb disk for the Digest files)

Target (Many-To-One scenario):
I'll have 4 different OS's (Win 2012/2012R2, Win 2016 and Win 2019) = 4 * 10Gb per OS = 40Gb

Total Global Cache Size = (# of Source WAN Accelerators) * (Size of Global Cache Configured in Target WAN Accelerator Properties) + Digest Size
So in my case = (2 source WAN accelerators * 40Gb Global Cache at target side since I have 4 OS's) + 30Gb (Digest size from Source 1) + (Digest size from Source 2) 120Gb = 230Gb in total.

But when I look at the Target disk it already consumed ~500Gb.

I'm sure i'm doing something wrong. But no idea what exactly...
Thanks for your help!
Maxim Karganov
Influencer
Posts: 22
Liked: 6 times
Joined: Jun 08, 2020 9:18 am
Contact:

Re: Wan Accelerator troubles/questions

Post by Maxim Karganov » 1 person likes this post

Hello,

Disk space for the Target WAN Global cache is pre-allocated at the beginning of a cache usage attempt.

According to the configured Target WAN cache size, B&R attempts allocating 400GB two times (2xSource WAN) and stops on 495GB because 1% of total space is always reserved.

Resizing the Target WAN Cache from 400GB to 40GB (we recommend 10 GB per every OS version on VMs) should help.
Argon
Influencer
Posts: 17
Liked: 1 time
Joined: Sep 29, 2015 1:17 pm
Contact:

Re: Wan Accelerator troubles/questions

Post by Argon »

Hello,

Aha, so the global cache size on the target server= x OS types * 10Gb which makes it in my case right now 40Gb
If I have 2 source WAN accelerators I don't have to multiply 40Gb * 2 and adapt the Cache size my own to 80Gb... That's where my mistake was I think.

So the Global Cache Size on the Target side is fine now. What about the Digest on the Target side?
They use this formula:
Total Global Cache Size = (# of Source WAN Accelerators) * (Size of Global Cache Configured in Target WAN Accelerator Properties) + Digest Size
Total Global Cache Size = (2 * 40Gb) + ***Digest Size

***What's the Digest size in my example if:
- Source 1 has a 30Gb Digest disk
- Source 2 has a 120Gb Digest disk

Will this make the following on the Target side:
Total Global Cache Size = (2 * 40Gb) + (120Gb+30Gb) = 230Gb in total.

Thanks for your advice!
Argon
Influencer
Posts: 17
Liked: 1 time
Joined: Sep 29, 2015 1:17 pm
Contact:

Re: Wan Accelerator troubles/questions

Post by Argon »

Perfect, all clear now!

Thanks!
Maxim Karganov
Influencer
Posts: 22
Liked: 6 times
Joined: Jun 08, 2020 9:18 am
Contact:

Re: Wan Accelerator troubles/questions

Post by Maxim Karganov »

Digests are not stored in a Target WAN cache. Thus, you don't need to explicitly add additional space to the configured Target WAN Cache for storing digests. It is sufficient to just keep that additional free space on the disk.

Let's take a look at the given scenario:
2% of 1.25Tb = 25Gb for the Source 1;
2% of 2.2Tb = 44Gb for the Source 2.

In this case, you need at least (25Gb + 44Gb = 69Gb) additional free space on the Target WAN disk for the digests.

Which in turn will be (2 * 40 + 69 = 149 Gb) total disk space for storing both cache and digests.
Remember to keep 40Gb in Target WAN settings.

Do not hesitate to ask if you have additional questions.
Post Reply

Who is online

Users browsing this forum: ante_704, Google [Bot] and 185 guests