Fixing zetaback erroring out when receiving FS is modified

[crazzy]

I've ran into this error again: http://lists.omniti.com/pipermail/zetaback-users/2013-July/000053.html

I digged a bit deeper into this and it seems that the issue here is that something on the backup server (which receives the stream) has modified the the filesystem. In our case I'll take a guess that it's a default Solaris cronjob (nfsfind).

It's a very simple patch, just adding -F to the zfs recv command which I've tested on two backup servers in our production environment with successful results, both on backup jobs which showed this error, and on other backup jobs not showing the error.

[crazzy]

Did another commit containing a fix for locking issues I've encountered on Solaris 11. Zetaback didn't even try taking backups of some datasets. After debugging with truss I found out the open() call on the directory failed with EISDIR. So in order to fix this I've made zetaback lock on a file instead. This has also been tested on the backup servers in our production environment with successful results.