NetBSD-Users archive

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

Re: Which file-system is good for power down?

On Dec,Saturday 29 2007, at 9:02 PM, Dieter wrote:

But our power is down every 1~2 week.

Ouch.  A UPS will help, but even with a UPS you can get a hang/crash
that corrupts data.  So a UPS is for riding through short outages,
it is NOT a guarantee that the machine never goes down hard.

Once power down occurs, the NetBSD server will check file system
at next startup, it will take as long as 10 minutes.

What you want is:

        soft updates
        disk's write cache turned off
        background fsck

This works great on FreeBSD.  No lost data.  No corrupted data.
Machine is available immediately after boot.  (Note that the
background fsck does use up some resources, such as disk i/o.)

Does NetBSD 4.0 have background fsck?

I have done a test on Linux (ext3), it take only 10 seconds to recover.

But Linux's ext will CORRUPT YOUR DATA!!! (I'll give you one guess how
I know this.)  Who cares how fast it boots if your data is corrupted?

How will ext3 corrupt my data after hard crash ?.

AFAIK soft-dependiences keep filesystem in consistent state with ordering metadata writes. softdeps keep ordered metadata(which were not written to fs yet.) in memory. And therefore if you have hard crash with heavy-io server you can loose more than with journaling fs. Keep number of softdeps in-memory buffers small is not solution because it slows your io then.

So, I think ffs v1 is not suitable for my environment.

FFS is fine.  Linux ext is crap that corrupts data.



Home | Main Index | Thread Index | Old Index