2 Oct
2009
2 Oct
'09
5:29 p.m.
Did anybody tried to install AVLinux with raid 5?Thank you
Giorgio
2 Oct
2 Oct
6:07 p.m.
On Friday 02 October 2009 21:29:41 Giorgio wrote:
Did anybody tried to install AVLinux with raid 5?Thank
you
I don't want to start a flame-war but all raids starting with f are considered
bad: http://www.baarf.com/
Use 10 or 01 or 1 if you want data-security by mirroring.
Or use LVM if you want big discs and mirrored partitions.
Arnold
6:19 p.m.
On Fri, 2 Oct 2009 22:07:40 +0200
Arnold Krille <arnold(a)arnoldarts.de> wrote:
On Friday 02 October 2009 21:29:41 Giorgio wrote:
You beat me to this, Arnold. The only thing I'd use RAID 5 for is an
array of disks where performance really doesn't matter to me, like a
disk-based backup server. I'd never use RAID 5 on a system where
performance mattered at all to me, like my studio machine!
--
======================================================================
Joe Hartley - UNIX/network Consultant - jh(a)brainiac.com
Without deviation from the norm, "progress" is not possible. - FZappa
Did anybody tried to install AVLinux with raid
5?Thank you
I don't want to start a flame-war but all raids starting with f are considered
bad: http://www.baarf.com/
6:27 p.m.
I know what you don't use now, what about telling me what you use? Do you
agree with arnold 10, 01 1?
On Fri, Oct 2, 2009 at 10:18 PM, Joe Hartley <jh(a)brainiac.com> wrote:
On Fri, 2 Oct 2009 22:07:40 +0200
Arnold Krille <arnold(a)arnoldarts.de> wrote:
On Friday 02 October 2009 21:29:41 Giorgio
wrote:
considered
You beat me to this, Arnold. The only thing I'd use RAID 5 for is an
array of disks where performance really doesn't matter to me, like a
disk-based backup server. I'd never use RAID 5 on a system where
performance mattered at all to me, like my studio machine!
--
======================================================================
Joe Hartley - UNIX/network Consultant - jh(a)brainiac.com
Without deviation from the norm, "progress" is not possible. - FZappa
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
Did anybody tried to install AVLinux with raid
5?Thank you
I don't want to start a flame-war but all raids starting with f are
6:44 p.m.
On Fri, 2 Oct 2009 22:27:17 +0200
Giorgio <anomalsound(a)gmail.com> wrote:
I know what you don't use now, what about telling
me what you use? Do you
agree with arnold 10, 01 1?
Right now I don't use any RAIDing or striping on my studio machine. I
have 3 hard drives and a CD-RW (it's still IDE-based). One drive's my
OS drive, one's my /tape drive with active sessions, and one's my
/tapebak drive with backups of active sessions and archives for older
sessions.
I have an rsync job that writes anything that's changed over to my
fileserver, which does use RAID 1+0 (mirrored and striped) to give me
both large filesystems and disk redundancy.
--
======================================================================
Joe Hartley - UNIX/network Consultant - jh(a)brainiac.com
Without deviation from the norm, "progress" is not possible. - FZappa
7:03 p.m.
On Fri, 2 Oct 2009 16:43:45 -0400
Joe Hartley <jh(a)brainiac.com> wrote:
On Fri, 2 Oct 2009 22:27:17 +0200
Giorgio <anomalsound(a)gmail.com> wrote:
My (highly limited) experience of RAID has been entirely negative - and
that's with out apparently using it :o
On two occasions now, I've tried to do what I thought was a perfectly
normal update on debian based systems which has resulted in something
called mdadm being upgraded, at which point it seems the entire
filesystem gets hosed :(
My policy these days is to make removal of mdadm the very first thing I
do after an install to a new machine.
--
Will J Godfrey
http://www.musically.me.uk
I know what you don't use now, what about
telling me what you use? Do you
agree with arnold 10, 01 1?
Right now I don't use any RAIDing or striping on my studio machine. I
have 3 hard drives and a CD-RW (it's still IDE-based). One drive's my
OS drive, one's my /tape drive with active sessions, and one's my
/tapebak drive with backups of active sessions and archives for older
sessions.
I have an rsync job that writes anything that's changed over to my
fileserver, which does use RAID 1+0 (mirrored and striped) to give me
both large filesystems and disk redundancy.
8:22 p.m.
On Friday 02 October 2009 22:18:55 Joe Hartley wrote:
On Fri, 2 Oct 2009 22:07:40 +0200
Arnold Krille <arnold(a)arnoldarts.de> wrote:
Especially on backup systems raid5 is very bad! The chances of a second disk
dying during recovery are much higher then with raid1(0) because all disks
have to be read to reconstruct the missing one. And as you probably don't have
all raid5 disks from different vendors/productions, the risk rises even more.
When you have two series of disks (each of the same branch/production) make
raid1 with one disk of each vendor and combine them with raid0 or lvm...
Arnold
On Friday 02 October 2009 21:29:41 Giorgio
wrote:
You beat me to this, Arnold. The only thing
I'd use RAID 5 for is an
array of disks where performance really doesn't matter to me, like a
disk-based backup server. I'd never use RAID 5 on a system where
performance mattered at all to me, like my studio machine! Did anybody tried to install AVLinux with raid
5?Thank you
I don't want to start a flame-war but all raids starting with f
are
considered bad: http://www.baarf.com/ 
3 Oct
3 Oct
3:20 a.m.
Joe Hartley wrote:
You beat me to this, Arnold. The only thing I'd
use RAID 5 for is an
array of disks where performance really doesn't matter to me, like a
disk-based backup server. I'd never use RAID 5 on a system where
performance mattered at all to me, like my studio machine!
RAID5 is not a good idea. Hard discs fail mostly due to environmental
causes, and if one drive fails, it is very likely that another will fail
soon, too. Probably while recalculating after replacing the failed disk
(or switching to the hot spare). You should at least use RAID6.
Flo
--
Machines can do the work, so people have time to think.
public key DA43FEF4 x-hkp://wwwkeys.eu.pgp.net
1:01 p.m.
New subject: RAID and AVLinux
You should at least use RAID6.
1. How does RAID6 differ from RAID5?
2. Is it actually possible to use software RAID or LVM mirroring with
AVLinux? I have not been able to find a way, short of installing to
non-RAID and converting the boot drive to RAID-1 using procedures I
found in a few places.
J.E.B.
5:36 p.m.
New subject: RAID and AVLinux
Jonathan E. Brickman wrote:
raid 6 uses a more elaborate algorithm than the simple parity of raid5,
to calculate two redundant bits. so you can recover from the failure of
two disks out of a raid6 array.
it's computationally more intensive, but the only feasible approach as
disk size increases, because, as other posters have pointed out, the
window of vulnerability during recovery is getting bigger and bigger.
my approach for music is this: all live recording goes to a mirrored
pair of disks. the data is then copied to a striped pair for editing,
with frequent backups.
at home it's actually just two disks with two partitions each, but since
i use either the striped or the mirrored part (except when copying stuff
over), it works pretty well, at low cost.
You should at
least use RAID6.
1. How does RAID6 differ from RAID5?
6:17 p.m.
New subject: RAID and AVLinux
Joern,
my approach for music is this: all live recording goes
to a mirrored
pair of disks. the data is then copied to a striped pair for editing,
with frequent backups.
Similar here: Recording on redundant machines with RAID1 each. Mixing is
done with the data on a SSD - seeking etc is a bliss thanks to the
extreme performance on nonsequential accesses.
Flo
--
Machines can do the work, so people have time to think.
public key DA43FEF4 x-hkp://wwwkeys.eu.pgp.net
6:44 p.m.
New subject: RAID and AVLinux
Unless I miss something on AVLinux:
On Saturday 03 October 2009 21:36:54 Jörn Nettingsmeier wrote:
Jonathan E. Brickman wrote:
raid 6 uses
a more elaborate algorithm than the simple parity of raid5,
to calculate two redundant bits. so you can recover from the failure of
two disks out of a raid6 array.
Loosing two disks means you got more then two coupled. If so, you have
definitely several of the same charge in that array. (Don't deny it, the human
laziness speaks against you.)
Recovering will still do heavy load on the disks. Now calculate the
probability of two disks of the same charge failing without a third disk of
that charge failing during recovery. Pretty low, huh?
Use raid10 with at least four disks or direct raid1 of several disks. Or stay
dynamic with lvm...
Good night,
Arnold
You
should at least use RAID6.
1. How does RAID6 differ from RAID5?
12 Oct
12 Oct
5:14 p.m.
New subject: RAID and AVLinux
Arnold Krille wrote:
Unless I miss something on AVLinux:
On Saturday 03 October 2009 21:36:54 Jörn Nettingsmeier wrote:
that remark is bogus. each disk design has a mean-time between failures
of X hours. that means of N disks, N/2 will have failed after time X for
a sufficiently high number of samples N.
the failure of each individual disk is a statistically independent
event. there is absolutely no synchronized self-destruct of disks
belonging to the same batch.
there is however the tendency of new disks to either die pretty fast or
last quite a while, and a very instructive long-term study by our
friends at google about aging disks. so if you're really careful with
your data, you don't wait for an old disk to fail but retire it after a
defined time. plus you should watch your smart data for early warning
signs. doing this makes large arrays quite manageable.
the problem these days is that disk sizes are getting ever larger.
that means more data is at risk, and the time to resync a degraded array
is also increasing. hence, raid5 might not be safe anymore. raid6 is
designed to not expose your data to a window of vulnerability at all -
one disk can always be swapped while the array remains fully redundant.
Jonathan E. Brickman wrote:
raid 6 uses
a more elaborate algorithm than the simple parity of raid5,
to calculate two redundant bits. so you can recover from the failure of
two disks out of a raid6 array.
Loosing two disks means you got more then two coupled. If so, you have
definitely several of the same charge in that array. (Don't deny it, the human
laziness speaks against you.)
Recovering will still do heavy load on the disks. Now calculate the
probability of two disks of the same charge failing without a third disk of
that charge failing during recovery. Pretty low, huh? You
should at least use RAID6.
1. How does RAID6 differ from RAID5?
6:14 p.m.
New subject: RAID and AVLinux
Joern,
the failure of each individual disk is a statistically
independent
event. there is absolutely no synchronized self-destruct of disks
belonging to the same batch.
It is not independent because the environment has an impact on the
failure rate (vibration, delta temperature etc.). So it is more likely
that another drive fails if one has already failed.
there is however the tendency of new disks to either
die pretty fast or
last quite a while, and a very instructive long-term study by our
friends at google about aging disks.
http://labs.google.com/papers/disk_failures.pdf
so if you're really careful with
your data, you don't wait for an old disk to fail but retire it after a
defined time. plus you should watch your smart data for early warning
signs. doing this makes large arrays quite manageable.
Performing checksum rebuilds on a regular basis is mandatory to detect
scan errors. On TB RAIDs with many discs they might me undetected for
months. I run a full rebuild at least once per month to detect scan
errors early.
the problem these days is that disk sizes are getting
ever larger.
that means more data is at risk, and the time to resync a degraded array
is also increasing. hence, raid5 might not be safe anymore.
RAID5 has never been safe. After one disc failed, you're caught with
your pants down.
raid6 is
designed to not expose your data to a window of vulnerability at all -
one disk can always be swapped while the array remains fully redundant.
Run RAID6 with at least one hot spare. Discs tend to fail when you're
not around.
The *real* problem is how to handle and archive the huge amounts of data
we work with today. Audio is a joke in comparison to video data. The
cheapest solution so far for mid term storage is still using more RAIDs,
unfortunately.
Flo
--
Machines can do the work, so people have time to think.
public key DA43FEF4 x-hkp://wwwkeys.eu.pgp.net
7:27 p.m.
New subject: RAID and AVLinux
On Mon, Oct 12, 2009 at 10:14:27PM +0200, Florian Faber wrote:
More likely is not the same as failure times clustering.
The latter will happen only in extreme cases, e.g. if your
server catches fire, or all disks get the same wrong power
voltages, etc. In such cases *all* disks will fail within
a short time, and there's little you can do about it.
In normal circumstances failure times are independent
random variables, even if their expected values are
scaled by the same ratio.
Ciao,
--
FA
Io lo dico sempre: l'Italia è troppo stretta e lunga.
the failure of
each individual disk is a statistically independent
event. there is absolutely no synchronized self-destruct of disks
belonging to the same batch.
It is not independent because the environment has an impact on the
failure rate (vibration, delta temperature etc.). So it is more likely
that another drive fails if one has already failed.
6:47 p.m.
New subject: RAID and AVLinux
Hi,
On Monday 12 October 2009 19:44:47 Jörn Nettingsmeier wrote:
Arnold Krille wrote:
Please listen to what I said: Two (or more) disks of the same manufacturer and
same type and maybe even continuous id-numbers.
And, bang!, now the failure of one device is highly connected to the failure
of the other one. Because the error (or call it small mis-fit) happening in the
production of the first is repeated in the second with a high probability.
And I worked on a manufacturing band several times, I know how humans catch
slight errors only after doing them the third or fourth time. By which its too
late to sort out and check the last three products.
Unless I miss something on AVLinux:
On Saturday 03 October 2009 21:36:54 Jörn Nettingsmeier wrote:
that remark is bogus.
each disk design has a mean-time between failures
of X hours. that means of N disks, N/2 will have failed after time X for
a sufficiently high number of samples N.
the failure of each individual disk is a statistically independent
event. Jonathan E. Brickman wrote:
Loosing two disks means you got more then two coupled. If so, you have
definitely several of the same charge in that array. (Don't deny it, the
human laziness speaks against you.)
Recovering will still do heavy load on the disks. Now calculate the
probability of two disks of the same charge failing without a third disk
of that charge failing during recovery. Pretty low, huh? > You should at least use RAID6.
1. How does RAID6 differ from RAID5?
raid 6 uses a more elaborate algorithm than
the simple parity of raid5,
to calculate two redundant bits. so you can recover from the failure of
two disks out of a raid6 array. there is absolutely no synchronized self-destruct of
disks
belonging to the same batch.
Yes, there is! I know from bad experience, both my own (not very tragic) and
that of a big company.
Have fun and make backups,
Arnold
5565
days inactive
5575
days old
linux-audio-user@lists.linuxaudio.org
16 comments
8 participants
participants (8)
-
Arnold Krille -
Florian Faber -
Folderol -
Fons Adriaensen -
Giorgio -
Joe Hartley -
Jonathan E. Brickman -
Jörn Nettingsmeier