What is the impact of using the "-nodata" option when running an "isync"?
For example, we had a partial copy of source data on a destination NetApp volume (a previously failed XCP copy due to stability issues on the source non-NetApp storage box. We had to use a different source box so couldn't reuse the previous copy catalog ID). We needed to establish a new catalog of the current state along with resuming the copy operation, so we opted to use isync rather than deleting all destination data and starting a fresh copy. (we're talking 250Tb over a 1Gb network link... ancient hardware with no option to improve the speed)
Due to the slow 1Gb link, iSync was taking a long time because XCP was reading in all data over the 1Gb link from the source so it could do a byte-for-byte check. Things drastically improved when we used the -nodata option, but what impact could using this have on the isync process besides being faster? What are we loosing out on?
The source data is live so changes can occur during each XCP run. Once the isync finishes I plan to use "sync" with the new catalog ID. Would this be ok?
Why does sync not have a -nodata option compared to isync? How does a sync check if a source file has changed without reading the whole file into the XCP server to do a byte-for-byte comparison?
Thanks in advance,
Paul.
