VMmark3: AuctionDB restore takes long time.

[Pavel2337]

I'm trying to run VMmark3 benchmark and it's currently at step "Wkld Setup" in process of restoring AuctionDB. It's been doing it for almost 24 hours. How long does it usually take? And how to check if it's something wrong? I have checked AuctionDB0 and ElasticDB0 VMs, they have disk and network activity. Here is screenshot:

Any help/suggestion would be greatly appreciated.

As of current status:

[dmorse]

Hi Pavel2337​,

We are aware this is an issue with VMmark 3.1, and our test/dev team is actively working on a fix.  I will update you when it is available.

Thank you for your patience.

David

[Pavel2337]

What is my option here?

1. Restart test run.

2. Clean, redeploy and restart test run.

3 Clean, download previous version, deploy and run test.

Also what previous working version should I get?

Thanks in advance,

[dmorse]

Unfortunately, the only option is to wait for an update to the VMmark 3.1 benchmark which will fix the issue you've reported.

[Pavel2337]

Can I run VMmark 3.0 instead? or it has the same problem?

Also after 41 hours test run stated but got DSWebA Abort and is trying to end iteration for 3.5 hours.

Should I just kill this test run?

[dmorse]

Correct, VMmark 3.0 and 3.1 have this problem.

My advice is to shut down all VMs and wait for a fix.  We are testing it as we speak.

Thanks,

David

[Pavel2337]

Thank you, David. Keep me posted on fix availability.

[dmorse]

This post has instructions for patching VMmark 3.1 to 3.1.1 which fixes the Weathervane issue you encountered:

VMmark 3.1 required patch to bring the benchmark to version 3.1.1