jcsd wrote:In SAN mode the backup proxy is a dual quad Intel E5506. I just started the job back up and checked the proxy after it started transfering and its running somewhere between 40-50-60-70-80%, I saw in hit the 90+% once. As the job slows down though the CPU usage drops way down.
I would be very interested in the CPU usage when the backup is slow.
jcsd wrote:
On the job I cancelled this morning it says source 2%, proxy 56%, network 99%, target 98%.
In a different instance of this job it says source 21% proxy 72% network 97% target 73%.
Interesting, so this would imply that the problem is all about the target storage. Have you monitored memory usage on the server during the backup, specifically lazy write cache and paging? I'm wondering if your having the same issues as
http://forums.veeam.com/viewtopic.php?f=2&t=8999 where backups start off fast, but as the write cache fills up and uses all of the memory the Windows machine slows to a crawl because it actually starts paging more and more. Normally this happens when the target is slower than the source.
Just to make sure I understand, when you are saying your destination is at a different site, are you using a repository at that site? How is the file being written to the remote site, via CIFS? What is the latency.
Have you tried simply copying a very large file?
Just looking at the information you have posted, it would seem that the storage at the remote site is "slow" (it may not be the storage itself, it may be the combination of protocol/latency if you're writing via CIFS). Please be specific as to the target storage and the how you are writing to it.
.