Unable to change replica Target Host

[Spad]

I've got a replica job that runs on a specific host with an NFS-attached datastore. This host is having various connectivity issues at the moment so we want to move the job to run on another host with the same NFS-attached datastore. Both hosts are running ESX v4.1.

When I try and change the target host in the job, I get the following error:

---------------------------
Veeam Backup and Replication
---------------------------
"esx1c" host uses different data transfer mode than the original host and cannot be selected.
---------------------------
Obviously I could create a new job, but I don't have the disk space to hold two sets of replicas for all the VMs, so I'd have to delete the existing replicas (unless you can import them into the new job somehow?).

[Vitaliy S.]

Hello Adam,

Could you tell me if you've specified service console credentials for your first host? If no, then you need to use "Force agentless mode" in the Data Transfer Tab of the second ESX host's properties.

Hope it helps!

[Spad]

Both hosts have service console credentials specified and they're both set to "Automatic" for data transfer engine.

[Vitaliy S.]

Judging by the error message the data transfer mode is different:

"esx1c" host uses different data transfer mode than the original host and cannot be selected.

Please try forcing agentless mode and see how it works for you. Thanks!

[Spad]

Nope, still getting the same message. I've also tried forcing service console agent mode and the same error is generated.

Thanks.

[Vitaliy S.]

Adam, in this case please send the corresponding job log to our support team, so they could advise you on the data transfer mode to use in order to continue replication job.

[Spad]

Vitaliy, just a quick note; I've tested this with some other jobs and they seem to be fine. I think it's probably happening because the original job was upgraded from 4.x when we moved to v5 and something hasn't "translated" correctly with regards to the transfer mode.

[Vitaliy S.]

Adam, thanks for the update, but still I would suggest comparing both job logs just to shed some light on your initial issue, though provided that you new jobs work fine, you may want to re-create this particular one.

[axs@moffatt.com]

I am having this issue. Ihad to remove and re-add my DR host becuase of an issue I had with the last VMware patch. I am running ESXi 4.1 with the latest patches.
I have several different replications defined and they are all having this issue. When I remove and re-add the replica as it states when I choose the storage (it does not understand that the storage is the same) I get this identical error.

[Vitaliy S.]

Hello Andrew,

Could you please give us more details on what you've done to that host? Have you re-added this host to a vCenter Server? What patches have you applied?

When I remove and re-add the replica as it states when I choose the storage (it does not understand that the storage is the same) I get this identical error.

Where do you remove and re-add replicas? Do you have the same message while starting a replication job?

Thanks!

[axs@moffatt.com]

I had to remove and re-add the host from the cluster because of an issue posted here:
http://www.veeam.com/forums/viewtopic.php?f=5&t=5366
I removed the host from virtual center and then re-added the host after a reboot.
That fixed the issue I was having but caused this new issue with my replication jobs.
The ESXi hosts were patched last Friday and all replication jobs worked this weekend (Friday and Saturday) so I don't think that is part of the issue. I started having this specific error
"host uses different data transfer mode than the original host and cannot be selected" when I removed and re-added the host.
The job logs showed that it could not find the replica files and when I checked the properties of the job it could not find the storage so I selected it and it instructed me to remove and re-add the VM replica (which I did). I then get the error and cannot continue.
Any help on this one would be greatly appreciated as all of my replication jobs failed last night.

[axs@moffatt.com]

Just got off the phone with support and I am being told that this is an issue because the host used to be ESX. I do not understand why this is an issue since no replications have failed since the upgrade (2 months ago). The issue only started to happen when I disconnected and reconected the host. It is being escalated to research and development at this point in the hopes that I will not need to rebuild all of my replication jobs. Of course I hope this is not the answer as these jobs will take the rest of the week to run the initial replication as I am replicating huge volumes.

[Vitaliy S.]

Andrew, could you please tell me your ticket number?

[axs@moffatt.com]

ID#544517

[ian0x0r]

I would be interested to know the outcome of this as I need to change my replication jobs to point from an ESX 4.0 host to an ESXi 4.1 host, but get the same error message.

[Vitaliy S.]

Ian,

Provided that you were using agentless mode to connect to your ESX host, then you should be able to change destination host from full blown ESX to ESXi host, but if you've specified service console crendetials (used SSH to connect to ESX host), then you should re-create your replicas from scratch.

The reason for this is that agentless and agent-based backup/replication modes use different methods in accessing the data (storage paths), and switching between those modes is not supported as per our Release Notes.

Thanks!

[ian0x0r]

Ah ok, thanks for the quick update!

Ian

[wolfgang.taitl]

The same issue here. Any news about it?

thanks!

[ian0x0r]

I had to recreate replication jobs :/

[Vitaliy S.]

Wolfgang,

Reading back through my response, I would recommend avoiding situations when you switch between agent-based and agentless modes, as in this case the only way to continue replication job would be starting this job from scratch.

Thanks!

[Bunce]

Is there a way to tell from the logs if agentless mode was used in a replication job, and does it cause a major difference in replica job performance?

Our replication speeds have dropped off considerably (compared to V4) since recreating our jobs in V5 (previously 600GB in 90 minutes - now 24 hours), however I've a gut feel if may not be due to Veeam but our simultaneous upgrade to ESX4.1 and the resultant SSH security change forcing agentless mode:
http://wahlnetwork.wordpress.com/2010/1 ... e-esx-4-1/

I'm putting together some details to submit a support rec, but from initial tests, our WAN optimation product (Riverbed) is still achieving the same exact reduction on Veeam traffic (approx 90%) - but is not receving anywhere near the throuhput.

So either we're not putting out the same throughput from the source, or it we can't receive that throuput at the destination end.

HotAdd mode with CBT enabled is indicated in the session details, using WAN block size. LAN backups jobs (to NFS store via SMB) of the exact same VM's takes about 50 minutes.

[Bunce]

Hmm - The WAN optimiser is receiving the traffic through port 902 whereas it used to be over 2500.

This post seems to indicate this occurs if running ESXi as a target, however we have ESX (full) at both ends.. Encryption is not enabled.

[Gostev]

Is there a way to tell from the logs if agentless mode was used in a replication job, and does it cause a major difference in replica job performance?

Yes / Yes

Hmm - The WAN optimiser is receiving the traffic through port 902 whereas it used to be over 2500.

This indicates that agentless mode is being used.

[mohammada]

[merged with existing discussion]

Hi Guys
I have several Replica Job running on 3 Different Host with shared storage . Need to shutdown two ESX and pont existing replicas to 3rd Host . so we want to move the job to run on another host with the same attached datastore.We are getting the following error :

""desx05" host uses different data transfer mode than the original host and cannot be selected.

We recently upgrade from Veeam replication 4.1 to V5 .
Can create new job but really dont want to do this as My volume of replication is Huge ( Around 7TB over Microwave Link) .So the Initial one will take very Long .

Please help me out if some one else face the same Issue . I can see a similer post on the forum but so solution .

Mohammad

[Vitaliy S.]

Mohammad,

Basically, the solution is not to change target host from ESX to ESXi host and vice versa. All the details are covered in this topic, please look through it.

Thanks!

[mohammada]

Hi Vitaliy
canot see a possible solution though from the above posts . Can you please point me out ?

[Vitaliy S.]

Mohammad, there is no solution. You shouldn't change the transferring method for an existing job. Here is the quote with explanation:

Ian,

Provided that you were using agentless mode to connect to your ESX host, then you should be able to change destination host from full blown ESX to ESXi host, but if you've specified service console crendetials (used SSH to connect to ESX host), then you should re-create your replicas from scratch.

The reason for this is that agentless and agent-based backup/replication modes use different methods in accessing the data (storage paths), and switching between those modes is not supported as per our Release Notes.

Thanks!

If this is not the case, then you need to contact our technical team directly.

[adapto]

[merged into existing discussion]

Hi I am tring to move a number of replica's from one server to another (changnig datastore as well)

To do this I have used the migrate option in vCenter to move the replica to the new server

I think this should essentially cover off the De-Register, Move the Files, Register requirement in the veeam notes.

However when I try and change the Veeam Replication Job I get an error message saying that the new host uses a different data transfer mode and cannot be used?

Anyone come across this or got any ideas what I may be doing wrong?

Thanks

Matt

[reddrinker]

Hi,

I am having the same problem, well I am about to, as I am in the process of moving first my DR then my production site's to ESXi 4 U1 from ESX 4.1. The bit i don't understand is how I can see/change from agent-less to agent and vice versa as I just connect to my Virtual Center Server from my Veeam Server and away it goes.

Secondly, as I am changing both ends, though only over a period of a few weeks, if i have to re-replicate it once i have chnaged over the dr site to ESXi will i then have to re-replicate again once I change the Production site to ESXi.

Is enabling remote tech mode the answer to this maybe, perhaps i will try it.

I really hope i don't have to re-replicate everything, and worse re-replicate it twice, oh, and loose my 21 days worth of replica's

Cheers

Jason

[Vitaliy S.]

Hi Jason,

If you haven't specified SSH credentials to enable Service Console connection, then you should be fine with switching your target ESX host to ESXi host. Enabling remote tech mode should make a trick, though we do not recommend doing it.

Secondly, as I am changing both ends, though only over a period of a few weeks, if i have to re-replicate it once i have chnaged over the dr site to ESXi will i then have to re-replicate again once I change the Production site to ESXi.

I'm pretty sure that changing source host connected through a vCenter Server would do not require re-replication.

Thanks!