Replication help

[GreenEnvy]

Hi all,
I'm confused reading the documentation for replication, so I thought I'd post here.
CaseID = 00680684

We want to replicate some data from Site 1 to Site 2. We have been doing this at the SAN level (Equallogic) but as we're replacing the Equallogic with a Nimble in Site 2, this is no longer an option, so switching to Veeam.

Background info:
Site 1 and 2 have a 30Mbps link between them, a few hundred miles apart.
Veeam is on version 7.0.871 in both sites

What we have:
In Site 1, we have a regular nightly backup of our file server, using reversed incremental. I copied the Veeam backup to an external hard disk, shipped it to Site 2, then imported it in to the Veeam backup repository on Site 2. I then restored it on Site 2, and added suffix "_Replica"

I tried creating a replication job on Site 1, with the following settings:
Destination: Cluster on Site 2
Datastore: a datastore on Site 2 (not same datastore I imported backups too)
Network: remapped VLAN's to match
Source and target proxy are autodetected
Repository for replica metadata is a repository at Site 1
Changed storage optimizations to WAN target
On Seeding, under "initial seeding", I expected to see the repositories in Site2 listed, but I only see repositories in Site 1. So I selected the store in Site1 that the original backup resides
For Replica Mapping, I enabled this and selected my original file server from site 1, and the restored backup at site 2

When I ran the job, it took 7.5 hours to calculate the digest for the first 55GB hard drive (OS drive), then 16 minutes for "Hard disk 1 55.0GB 54.1GB Read at 50MB/s [CBT]"
Next hard drive is 1 TB, it took 93 hours to calculate digest, but on the next step it failed after about 9 hours saying it read 94.1GB of 1.0TB at 3MB/s.

Code: Select all

Error was: 11/16/2014 8:24:46 PM :: Error: Client error: Failed to open VDDK disk [[GRNimble1] BU-SRV-FS01_replicated/BU-SRV-FS01_1-000002.vmdk] ( is read-only mode - [false] )
Failed to open VMDK.
Logon attempt with parameters [VC/ESX: [grsrv-vcenter0];Port: 443;Login: [****];VMX Spec: [moref=vm-22297];Snapshot mor: [snapshot-22357];Transports: [nbd];Read Only: [false]] failed because of the following errors:
Failed to process [srcReplicateVddkDiskContent] command.
An existing connection was forcibly closed by the remote host

If anyone can tell me what went wrong that would be great. Also why the calculating digest info took so long. Is that normal or a configuration issue?

Whole log file here:

[EDIT] log has been removed by moderator

[Vitaliy S.]

Hi Lee,

Digest calculation is required to detect all the changes happened on the source VM and then compare this data with a target VM.

On Seeding, under "initial seeding", I expected to see the repositories in Site2 listed, but I only see repositories in Site 1. So I selected the store in Site1 that the original backup resides
For Replica Mapping, I enabled this and selected my original file server from site 1, and the restored backup at site 2

Not sure I fully understand what you're trying to do here. You should either seed or map your replication jobs. When you select datastore located on the source site (as a seeding option), Veeam backup server will go and restore your VM first and then do mapping procedure automatically. Since you've already restored the VM on the target site and have two proxy servers installed, please try to just map your replication job and check out its performance.

Thanks!

[GreenEnvy]

Ok,
Can someone tell me if I am doing this right?
Some things have changed since I started this:

Site A:
Vsphere 5.5
Veeam 8.0 917
Vsphere is installed on a Windows 2012 guest VM, Veeam is installed on the same guest
Nimble array holding all the VM's

Site B:
Vsphere 5.5
Veeam 8.0 917
Vsphere is installed on a Windows 2012 guest VM, Veeam is installed on the same guest
Equallogic array holding all the VM's.

Both sites are production sites. I'd like to setup replication so I can have copies of some VM's from site A be stored at site B
So I loaded a copy of the backup files from site A (VBK and VBM file) onto an external hard disk, and shipped to Site B.

At Site B, I copied all the files into the folder where Site B's backups are located (so it's in same backup repository), then edited the backup repository and had it import data. I can now see the guest OS's at Site B in Veeam, showing up with the correct repository.

At Site B, I then created a replication job, Selected the guest VM from Site A on the 'Virtual Machines' screen.
For Destination, I selected the cluster at Site B, and the datastore I want it to reside on at site B
On Data transfer, I left both on Automatic selection, and direct (don't have plus)
For Seeding, I checked the box under initial seeding, 'get seed from the following backup repository', and selected the repository at Site B that I imported Site A's backup into
Clicked Finish.

When I run the job, it first processes 'myVmName [seed]', which runs fine.
Then it tries to process 'MyVmName' but errors out. The error says 'Error: File does not exist or locked. VMFS Path: [[GRNimble1]MyVmName/MyVmName.vmx]
Please, try to download specified file using connection to the ESX server where the VM registered.
Failed to create NFC download stream. NFC path: [nfc://conn:grsrv-vcenter0,nfchost:host-14981,stg:datastore-15419@MyVMName/MyVmName.vmx].


So first off, have I setup the job in the right direction? I setup the replication job to run from Site B, is that correct?
Second, it looks like Veeam is trying to access the datastore at Site A through the vCenter at Site A, correct?
At site A, I have gone into ISCSI initiator and connected to the Nimble volume.

I also tried it a different way, restoring "MyVmName" at Site B, then setting up a replication at Site B using the restored VM instead of the seeding from repository option. That failed with the same error, file does not exist or locked.

Any thoughts?

[foggy]

Looks like you're doing everything right. Please review this thread discussing similar issue and contact support directly if none of the hints help.

[GreenEnvy]

As a followup for anyone else with this issue, it was a routing issue.

Our Vcenter servers have two virtual NIC's in them. There is one on the "server" vlan, and one on the "iscsi" vlan. Those both were routable to the equivalent one in the other office, however they weren't routable across types between the offices.
So Server VLAN in site A could talk to Server VLAN in site B. ISCSI vlan in site A could talk to iSCSI vlan in site B.
However Server VLAN in site A could not talk to iSCSI vlan in site B.

We didn't think this would be needed, as we thought Veeam/vcenter would know to keep that traffic separate, but apparently not.
Added some routes for this and now replication is working fine.