Veeam Replication and WAN accelerator disk usage

[Lex]

Hi,

Trying to get Veeam 8 replication working with WAN accelerators. After several weeks of testing and testing i come to the conclusion that the manual on this part needs to be, well, re-written.

Let me put the picture right, we have a source enviroment with 2.3Tb of vm's.
To get it all working, we needed on the target side a WAN accelerator cache which can hold the 2.3Tb in cache?!?!?!?!?! Smaller cache results in error: not eneugh diskspace available when building the digest. Thus the amount of space needed on the target side is the amount of the backup seed, in our case 900Gb + the space in use by the VMs 2,3Tb + some extra overhead produced by the Veeam software.

The manual speaks of using SSD which will result in a faster result. This isn't the case sinds the bandwith within the software is somewhere limit to 1Gb while 10Gb is available and even direct attached SAS 12Gb disks versus iscsi does not improve in time. 550Gb will take 1,5 hours for digest and finger printing will take the same amount of time.

I could be wrong and there could be a setup failure, but as support has allready been looking at our setup and came up with nothing other then reboot. I presume then that the setup is correct.

Case id: 00693201

[Gostev]

Let me put the picture right, we have a source enviroment with 2.3Tb of vm's.
To get it all working, we needed on the target side a WAN accelerator cache which can hold the 2.3Tb in cache?!?!?!?!?!

No such requirement. You will be fine with even 20-30GB of cache on the target side WAN accelerator.

Smaller cache results in error: not enough diskspace available when building the digest.

This needs to be investigated, but I believe this error does not talk about target WAN accelerator cache size, and rather either source WAN, or metadata repository.

The manual speaks of using SSD which will result in a faster result. This isn't the case sinds the bandwith within the software is somewhere limit to 1Gb while 10Gb is available and even direct attached SAS 12Gb disks versus iscsi does not improve in time. 550Gb will take 1,5 hours for digest and finger printing will take the same amount of time.

You observation is correct, this SSD requirement is coming from the previous version (v7). It's not as essential to use SSD for cache in v8 in most scenarios. We will adjust the requirements later, based on observing v8 perform in real-world environments. Meanwhile we are going with "safe" defaults with our recommendations to the users.

I could be wrong and there could be a setup failure, but as support has allready been looking at our setup and came up with nothing other then reboot. I presume then that the setup is correct.

No, this response only means that you are dealing with fresh T1 support engineer and nothing else wait for your case to be escalate to a higher tiers for some actual investigations.

[Lex]

On the target site e:\VeeamWAN is the space, source has got c:\VeeamWAN
Only when i made the space above 2.3Tb the job does work.

[Lex]

At this point i'm looking for more then a hour at this screen, from what i can see it is doing something on the target at 20Mb.

[Vitaliy S.]

Hi Lex, please avoid posting log snippets as it is advised by our forum rules when you click New Topic. The last operation is checking the source and target data that has been transferred. As to the error message with the free space, please continue working with our support team. Thanks!

[Lex]

Sure Vitaliy S., no problem.

But could it then escalate to the right level. As stated at the start of this thread, the manual does not reflect the real deal and the job is now running for more then 24 hours and consuming over 100% more diskspace then expected.

With regards,

Lex

[Lex]

Made a new case for this: 00716790

[Vitaliy S.]

Lex, you can always ask your current engineer to escalate the case to a higher tier engineer, no need to open a new case.