ZFS vfs.zfs.cache_flush_disable and ZIL reliability

[Ivan Voras]

On 17/03/2011 08:45, Jeremy Chadwick wrote:

> General question to users and/or developers:
>
> Can someone please explain to me why people are so horribly focused (I
> would go as far to say OCD) on this topic?

For me, it's a matter of statistics; if this can improve the chances of 
data surviving by (just guessing here) 10%, it might be worth it.

> Won't there *always* be some degree of potential loss of data in the
> above two circumstances?  Shouldn't the concern be less about "how much
> data just got lost" and more about "is the filesystem actually usable
> and clean/correct?"  (ZFS implements the latter two assuming you're
> using mirror or raidz).

As an admin, I'd much rather have a file system that's clean with some 
data lost, but as a user I think I would be unhappy with any data loss 
:) Backups still rule.

ZFS is covered (presumably, as I don't really get from the code how it's 
supposed to work - any clarification for pjd@, mm@ and others would be 
appreciated) and I've talked with McKusick about BIO_FLUSH in UFS and it 
will happen as soon as I have the time to follow up.