I have 2 CentOS systems, one is rsyncing a ~411 gig directory to another.
On the receiving system, I typically run du -cs * |grep total;sleep 5
as a way to monitor the background rsync operation and make sure it is actually happening, since I don't know of a good way to monitor it as it runs otherwise.
The behavior I'm seeing is that du will show the total number of kilobytes go up, then go back down to a previous (and exact) value:
430235952 total
430235952 total
430235952 total
430235952 total
430235952 total
430235952 total
430235952 total
430235952 total
430498032 total <----
432333040 total <----
434430192 total <----
430235952 total <---- Back down to original value>
430235952 total
430235952 total
430235952 total
430235952 total
430235952 total
431284468 total <---- Up again
434430196 total <----
430235952 total <---- Back down again>
430235952 total
430235952 total
430235952 total
431284468 total <---- ???
434430196 total <---- ???
430235952 total
430235952 total
430235952 total
Does anyone know what's going on here?
I could be wrong but it looks like maybe a file is being transferred and it is later found to have corrupt data so rsync deletes the corrupt copy and tries again, you could run rsync with the -v flag so it outputs verbose information but you really shouldn't sweat it too much. If there is a problem with the transfer it will likely output warning information letting you know what's going on and not just hang indefinitely.