[Top] [All Lists]

Re: restoring after running xfs_repair

To: James Rich <james@xxxxxxxxxxxxx>
Subject: Re: restoring after running xfs_repair
From: Michael Sinz <msinz@xxxxxxxxx>
Date: Fri, 27 Sep 2002 14:03:39 -0400
Cc: Eric Sandeen <sandeen@xxxxxxx>, XFS mailing list <linux-xfs@xxxxxxxxxxx>
References: <Pine.LNX.4.44.0209261904460.2090-100000@xxxxxxxxxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.1b) Gecko/20020813
James Rich wrote:
On Thu, 26 Sep 2002, Michael Sinz wrote:

This sounds like the IDE-write cache does not flush before power off
problem.  There is a "fix" to this in some systems that has the shutdown
process, after syncing/unmounting stuff do a powerdown of the IDE
drives to force them to write their caches before doing a shutdown
of the ATX power supply via APM/APCI.

Yes, this seems likely now.  After disabling ACPI and APM I noticed a
message indicating that the IDE drives are being flushed.  I don't think
this message was there before.  Probably a problem with my dist. (I'm
using Slackware - I'll report it there).

Note that a message that says they are being flush does not mean that
they actually are - many (most) current IDE drives don't actually flush
or sync when you tell them unless you tell them to spin down/power off.
(At the drive level)

This is why many distros are now not only doing the flush/sync command
to the drives but then also doing a powerdown to the drives before
hitting the ATX power (APM/ACPI).  Many others has significant loss
due to the power supply turning off faster than the drives had a chance
to write their 1M/2M/8M buffers back to disk.

Michael Sinz -- Director, Systems Engineering -- Worldgate Communications
A master's secrets are only as good as
        the master's ability to explain them to others.

<Prev in Thread] Current Thread [Next in Thread>