On Wednesday 09 December 2009, nixlists nixlists wrote:
> Hi. My 'softraid' mirror is not being detected and assembled at the
> boot time. I must run 'bioctl' to assemble it after a reboot. This
> started happening after I removed another softraid mirror from the box
> (physically - the card and th
Am 10 Dec 2009 um 23:00 schrieb Marco Peereboom:
On Thu, Dec 10, 2009 at 05:00:34PM -0500, nixlists wrote:
Hmmm. I've used hardware raid cards for mirrors that have the
verify function.
It would be interesting to know how and what those cards do.
They read the data to make sure the disk is
On Thu, Dec 10, 2009 at 6:41 PM, Marco Peereboom wrote:
> probably a crappy card or disks.
3ware Escalade 8006-2LP :(. I know - not well supported because 3ware
are the M$ of RAID.
On Thu, Dec 10, 2009 at 06:17:24PM -0500, nixlists wrote:
> On Thu, Dec 10, 2009 at 6:00 PM, Marco Peereboom wrote:
> >> Does this mean there's little advantage of hardware mirror raid over
> software?
> >> So software mirror raid increases chances of data corruption while
> decreasing
> >> the ch
On Thu, Dec 10, 2009 at 6:00 PM, Marco Peereboom wrote:
>> Does this mean there's little advantage of hardware mirror raid over
software?
>> So software mirror raid increases chances of data corruption while
decreasing
>> the chances of downtime. True for hardware as well?
>
> There are pro and co
On Thu, Dec 10, 2009 at 05:00:34PM -0500, nixlists wrote:
> On Thu, Dec 10, 2009 at 3:37 PM, Marco Peereboom wrote:
> >> So in what cases does softraid degrade the mirror then, other than
> >> pulling the disk out?
> >
> > When an I/O fails.
> >
> >> How is hardware mirror raid different?
> >
> >
On Thu, Dec 10, 2009 at 3:37 PM, Marco Peereboom wrote:
>> So in what cases does softraid degrade the mirror then, other than
>> pulling the disk out?
>
> When an I/O fails.
>
>> How is hardware mirror raid different?
>
> It isn't.
>
>>
>> Thanks.
Does this mean there's little advantage of hardwa
On Thu, Dec 10, 2009 at 03:21:46PM -0500, nixlists wrote:
> On Thu, Dec 10, 2009 at 10:12 AM, Marco Peereboom wrote:
> >> So softraid can't detect if the data is written differently to the
> >> drives? In what sort of cases would one expect the mirror to become
> >> corrupt? Kernel crash? Hardware
On Thu, Dec 10, 2009 at 10:12 AM, Marco Peereboom wrote:
>> So softraid can't detect if the data is written differently to the
>> drives? In what sort of cases would one expect the mirror to become
>> corrupt? Kernel crash? Hardware crash? Does softraid detect this? What
>> failures does it detect
On Thu, Dec 10, 2009 at 12:39:29AM -0500, nixlists wrote:
> On Wed, Dec 9, 2009 at 11:54 PM, Marco Peereboom wrote:
> >> What happens if the written data is not the same for some reason (and
> >> what's the likelihood of this happening)? How does the OS and fsck
> >> behave in this case?
> >
> > T
On Wed, Dec 9, 2009 at 11:54 PM, Marco Peereboom wrote:
>> What happens if the written data is not the same for some reason (and
>> what's the likelihood of this happening)? How does the OS and fsck
>> behave in this case?
>
> There is no way to quantify that. You are the one that needs to
> calc
On Wed, Dec 09, 2009 at 11:10:35PM -0500, nixlists wrote:
> On Wed, Dec 9, 2009 at 10:42 PM, Marco Peereboom wrote:
> > It isn't identical because softraid does not initialize raid 1 because
> > it doesn't matter. So all written data is the same but areas where the
> > volume has not been written
On Wed, Dec 9, 2009 at 10:42 PM, Marco Peereboom wrote:
> It isn't identical because softraid does not initialize raid 1 because
> it doesn't matter. So all written data is the same but areas where the
> volume has not been written before will vary on the individual disks.
What happens if the wr
On Wed, Dec 09, 2009 at 06:31:43PM -0500, nixlists wrote:
> Also if I am paranoid about mirror data being exactly the same on the
> two halves (yes, I understand softraid should guarantee it, but
> still...), how can I verify it? Or this functionality currently
It isn't identical because softraid
Also if I am paranoid about mirror data being exactly the same on the
two halves (yes, I understand softraid should guarantee it, but
still...), how can I verify it? Or this functionality currently
nonexistent? Or am I asking a stupid question because softraid is
guaranteed to notice these things a
On Wed, Dec 9, 2009 at 4:30 PM, Marco Peereboom wrote:
> jsing is working on a "add auto assemble flag back" button. For now you
> are stuck with bioctl -c until that is done.
>> 'softraid0 at root'
>>
>> dmesg shows that softraid is not complaining at all, just the standard
>> 'softraid0 at root
jsing is working on a "add auto assemble flag back" button. For now you
are stuck with bioctl -c until that is done.
On Wed, Dec 09, 2009 at 04:21:23PM -0500, nixlists wrote:
> On Wed, Dec 9, 2009 at 9:58 AM, Marco Peereboom wrote:
> >
> > I think you mean assemble instead of build. If I follow
On Wed, Dec 9, 2009 at 9:58 AM, Marco Peereboom wrote:
>
> I think you mean assemble instead of build. If I follow your meager
> description of the issue correctly this should work. You can move a
> softraid volume to another machine and it should auto assemble. The
> trick is to have all piece
I think you mean assemble instead of build. If I follow your meager
description of the issue correctly this should work. You can move a
softraid volume to another machine and it should auto assemble. The
trick is to have all pieces in good shape. A dmesg might help because a
disk that wasn't au
19 matches
Mail list logo
