I've been suffering from what appears to be sporadic conversion performance (ranging from 800Kb to 50MB/s), I've ran two conversions side by side of windows 2003 servers (yes eol, yes no support, but this is what our business demands) both in the same source location and both going to the same destination infrastructure - One will run at around 20MB/s and the other around 4MB/s. If all my migrations ran at 4MB/s i'd wonder whether its infrastructure, but it seems to be specific to servers therefore i would like to focus on this example and see if theres anything from server level that can be done.
Actions I have taken.
To summarise in one question: what in addition to the above can be done at server level to increase performance.
a few clarifications first:
"SSL set to false." - I assume this means "<useSsl>false</useSsl>" in worker config as in Increasing the cloning performance, correct?
"Ensured NIC speed set" don't know what you mean by that, hope it's ok.
Other factors that affect performance might be
cloning level - file level is slower than block level; file level will be triggered by shrinking the volume or changing the cluster size
ESX host performance. The conversion establishes a connection between the source machine and the destination host and uses the proprietary NFC protocol. I.e. the host current load, disk speed etc. will affect the performance, too.