Just to add a little bit of closure to my portion of the thread.
Ultimately, the only way we could get a good backup of our largest VM's was to use SAN or Network transport mode on a physical proxy server.
We are in the process of combing through our VM infrastructure to find out what would impact this and are working with support on why our virtual proxy server seems to be having trouble.
Very interesting. So a virtual proxy failed with hotadd and network mode, but the physical machine worked with SAN or network mode? Was there any significant other differences between the two (amount of RAM, number of cores/vCPUs, OS version, antivirus, etc.) What about host resources, was the memory reserved on the host? Once or twice I've seen an issue where memory usage on the proxy would cause swapping to occur on the ESXi hosts due to memory pressure and performance would tank. Please keep us updated if you make progress with the virtual proxy.
The virtual proxy failed with SAN mode too. For testing I placed the proxy server all on a host by itself with 12gb of RAM (96 on the host total) and it still occurred. This was a fresh Win2k8 R2 install too so I am very perplexed.
SAN mode on the physical server also suffered a significant performance drop about 1/2 through the job and where as the first 15 VM's or so had source as the bottleneck, the last 15 had target listed. This was all despite the fact there were 2gb free out of 6gb on the physical proxy and none of the 6 cores were ever more than 90% utilizied.
I will update this thread as we determine what is going on.
I am facing the same issue with a replication job, where 2 VM (biggest of the pool, 2/3To each) are failing with the same error, I already tried the solutions given in this thread, without luck.
Could someone give an update ?
Cedric, there is another existing topic regarding the same error, please review for possible tips. However, the most effective way to address it would be contacting our technical support team directly. Thanks.