by Borsh » Thu May 16, 2013 1:10 pm people like this post
Hello! We have two site: A and B. In each of them there are two hyper-v 2008 r2 clusters with 6 CSV. Veeam B&R 220.127.116.11 (with patch 3). After patch is installed on veeam server all hosts were rebooted. Onhost backup proxy and MS VSS hardwire provider is used. We start multiple replication jobs and after their completion, Hyper-V hosts lost connection to the CSV volume, on which the replicated virtual machine are located. Event ID in MS event viewer on hosts: 1584, 5120. MS Hotfix KB2637197 has already installed.
by v.Eremin » Fri May 17, 2013 8:51 am people like this post
Under certain circumstances you might get the said errors due to:
• Lack of hard disk space required for snapshot allocation. It’s recommended to have at least 20% free space on storage. • Lack of hardware VSS provider. Or it has to do with hardware provider working or being installed wrongly.
Additionally, has our support team already commented on this issue or the case is still under investigation?
by Borsh » Fri May 17, 2013 10:52 am people like this post
Hello, Vladimir! Yes, VEEAM support has already contacted with me. I think it's lack of hardware VSS provider, because we have enough disk space (more than 70%). In our case, MS VSS hardwire provider is using. Can i use another VSS hardwire provider if backup proxies are on hyper-v hosts? For example, our SAN storage is HDS AMS 2300, so HDS VSS provider.
by Vitaliy S. » Fri May 17, 2013 2:32 pm people like this post
Snapshot will be taken on your Hyper-V server and based on the configuration, it will either be transferred to an offhost proxy or will remain on the production Hyper-V server. Please take a look at our User Guide (pages 12-13) for further reading.
We have the following setup: Hyper V Cluster OS: Windows Server 2008 R2 Servers: 2xHP Proliant DL380 G7 Storage: HP P2000 G3 Veeam: V 6.5 Patch 3
We want to have reverse-incremental backups of our VMs. Issue is that as long as all VMs are on a single host (host 1 or host 2), the veeam backup does not fail but if we distribute the VMs between host 1 and 2 then intermittently during the backup, one of host randonmnly will loose connection to eth CSV with typically errors 5120 and 5142. We have to restart the affected host and only then it rejoins in the Cluster
by Borsh » Tue May 28, 2013 9:25 am 1 person likes this post
As a result, the problem has been that Jumbo frame was not enabled in the Cisco kernel switch. After it's was enabled - hosts does not lose connection to CSV when it"s used MS software provider or Hitach hardware provider.