tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: WAPBL and write cacheing

Edgar Fuß <> writes:

> May I repeat a question I asked a month ago?
> I learned that it's safe to use write cacheing on a SATA drive with FFS+WAPBL.
> I would like to know
> 1. whether it's safe on SAS drives, too
> 2. whether it's safe with plain FFS, too

I'm not 100% clear on this, but the concept is that wapbl invokes a
cache flush on the drive when committing the log.  So if the flush
really works on SAS, it should be the same as SATA.  The cache flush
does happen with umass and a sata USB2 box; I have a drive that takes
300 ms to flush caches (WD Elements), so I did notice.

I would say plain FFS is only ok with write cacheing if you have stable
power.  Which means not really, but everybody does it anyway :-)

There are sysctls:

vfs.wapbl.flush_disk_cache = 1
vfs.wapbl.verbose_commit = 0

and you could modify the verbose_commit, if it doesn't already do it, to
print the time it took to execute the flush.

Attachment: pgptK4EJgRBJC.pgp
Description: PGP signature

Home | Main Index | Thread Index | Old Index