Hello,
I've got rsync --daemon running on an OSX machine, and I have it exporting
the entire hard disk as [rootpath]. On my Linux machine, I run this:
rsync -avvv --delete --exclude /Volumes macbox::rootpath /mnt/bkup/macbox
I redirect the output from that command to a file (nohup.out), and then tail
the file, and I see that it never gets beyond the recv_file_name() stage.
In about 4 hours, the nohup.out file is 4.2GB (yes, gigabytes), and the
rsync process dies with:
rsync: connection unexpectedly closed (847940065 bytes read so far)
rsync error: error in rsync protocol data stream (code 12) at io.c(342)
_exit_cleanup(code=12, file=io.c, line=342): about to call exit(12)
...and it hadn't transferred a single file. There's only about 35GB of
data
on the macbox, and if I run a bunch of individual rsync commands to copy all
macbox's directories one at a time, they all complete fine.
So I run the all-at-once rsync command again, but this time I kill it after
10 minutes to check the nohup.out file (which is now about 170MB), and it
turns out that it's full of duplicate lines:
[me@linuxbox]$ cat nohup.out |sort|uniq|wc
469433 1016118 53240522
[me@linuxbox]$ wc nohup.out
1591372 3347600 179505609
So that's 1.6 million recv_file_name() lines, 70% of which are duplicates.
It appears that it just keeps receiving the same filenames over and over
till it craps out. I've done this a few times, rebooted the machines, etc,
and it happens every time.
Any ideas about this?
Thanks,
Anthony DiSante
http://nodivisions.com/