Poor SOBR offload performance - poor support response.

[ConradGoodman]

Support case raised over a week ago Case # 04461796.

We are experiencing very long waits offloading blocks to AWS.

Single SOBR with single local extent and single capacity tier extent.

Daily copy job set to mirroring mode, around 60VMs.

I am starting to lose confidence in this software, we are struggling to get deltas uploaded because the blocks are waiting on the veeam server for many periods of 40-60 mins before it even attempts to offload them.

So far support has been terrible taking a long time to respond, and coming up with little or no answers.

This is a small implementation, but I need to chuck a lot more data at it (SQL backups yet to be added).

At the moment, its taking so long that we will not get the backups offloaded within 24 hrs if I add my SQL backups.

We've paid a lot of money and put faith in veeam with this solution. I cannot believe this problem is that difficult to solve, you must have customers with 1000's of VMs going into AWS.

[veremin]

Actually, from the case history it would appear that support has been working with your very closely, as opposed to "poor support response".

I see your case was taken over by an escalation engineer several days ago. The first potential fix was already provided to you yesterday, after reviewing your configuration database. As I can see, the fix didn't make any difference, which is what you have notified the support engineer several hours ago. Their have then continued investigation with R&D, so please wait for the update on the next steps.

Thanks!

[soncscy]

Not talk down to you, Conrad, but I work with clients with many thousands of VMs to offload, and it's never been an issue.

With decent servers behind the copies and a nice connection, this has been mostly seamless for my clients.

The ones I had issues with, they overestimated or just weren't straightforward about the server resources or the uplink.

I am not accusing you of either, but there is a lot of ambiguity with your message just from my point of view. Details that are required, imho, to properly assess the validity of your concern.

But, that's just my take, but I do think you'll get more info here with more details.

[ConradGoodman]

It is worth sharing here, that our Veeam environment is in a secure configuration. We use a proxy server to offload to AWS.

This is configured in WinHTTP proxy settings with netsh, as per the veeam documentation.

There is a bug in veeam, in which the part of its code that checks for updates from http://dev.veeam.com for AWS bucket locations, does NOT respect WinHTTP proxy setting.

The fix from Veeam was to use pstools to execute iexplore as LocalSystem, and set the proxy setting for dev.veeam.com.

This improved backup performance by ~ 1000%!

I would like to hear of an update that fixes this issue properly. There must be many people running proxies in secure environments.

[Gostev]

Good news, v11 uses HTTPS for this.

[ConradGoodman]

Does it respect the WinHTTP proxy setting?

[Gostev]

No - this is not a simple fix, so it's too late to address in v11 at this point. Plus, we only had a single support case in all this time anyway. However, we did address the impact of this issue on the actual offload functionality, which was what prompted this topic.