|
An update:
> 1. The server-side daemon is handled in user-space. This is
> slow, and causes other problems -- for example, the buffer cache
> doesn't get updated when you change the underlying disk directly.
> Not good :)
After some more tinkering with nbd, I've found that the buffer cache
problem only shows up if the exporting server writes to the filesystem,
but if the server receiving the export writes, it works just fine.
Since I intended to mount the drive read-only on the exporting server
anyway (per #2), this is no longer a problem for me.
Next, I will tinker with including an exported device in a RAID1 array.
I'll get back to you if/when I have something to report :)
(It's still slow, but hey, slow is better than nothing, right?)
Thanks,
Kyle Sparger
|
