4.x device failure?

[Scott Long]

It's not impossible to remove a disk object from a running system,
you just have to be very careful about it.  Like I mentioned before,
keep a reference count on your disk, and only call disk_destroy() when
it has gone to 0 (and prevent new opens while you are waiting to
destroy).  This isn't great, of course, since you can't destroy a disk
with active references, but it's not impossible.

Scott

Sam wrote:
> Ah.  Well that answers it.  I'll fail unloading of the
> module if devices are open.
> 
> This leads to a curiosity as to how the RAID
> failover modules work at all.  I guess if you abstract
> the real disk(s) under a logical disk you can change
> the bottom layer without affecting the top.  I'll have
> to look at that code a little closer.
> 
> Thanks -
> 
> Sam
> 
> On Mon, 13 Sep 2004, Poul-Henning Kamp wrote:
> 
>> In message <Pine.LNX.4.60.0409131202270.6275 at athena>, Sam writes:
>>
>>>
>>
>>> Surely there's a way to pull a disk out from under
>>> those that have it open without a panic?  Opinions?
>>
>>
>> No, we're not quite there yet.  I/O errors from disk devices very
>> often leads to filesystem or buffer cache panics.  We're working
>> on it.
>>
>> -- 
>> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
>> phk at FreeBSD.ORG         | TCP/IP since RFC 956
>> FreeBSD committer       | BSD since 4.3-tahoe
>> Never attribute to malice what can adequately be explained by 
>> incompetence.
>> _______________________________________________
>> freebsd-arch at freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-arch
>> To unsubscribe, send any mail to "freebsd-arch-unsubscribe at freebsd.org"
>>
> _______________________________________________
> freebsd-arch at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-arch
> To unsubscribe, send any mail to "freebsd-arch-unsubscribe at freebsd.org"