7 Oct
2012
7 Oct
'12
8:53 p.m.
Message: 19
Date: Sun, 07 Oct 2012 03:51:12 -0600
From: Eric Smith <eric at brouhaha.com>
To: General Discussion: On-Topic and Off-Topic Posts
<cctalk at classiccmp.org>
Subject: Re: disk flaws, classic vs. modern
Message-ID: <50715090.2020207 at brouhaha.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
On 10/07/2012 01:38 AM, jim s wrote:
recently I had a system with a 1.5tb seagate grow a
count of
"uncorrectable offline sector count" errors. [...] To complicate
things a bit this was part of a LVM raid ext3 raid 5 set,
When you have a drive go bad in a RAID 5, it's best to pull the drive,
put a new one in its place, and start a rebuild. That's the point of
using RAID 5. Trying to recover data from the failing drive is mostly a
waste of time. Of course, if the drive wasn't in a RAID 5, mirror,
etc., you wouldn't have that option.
My experience is that once it starts it just gets worse, especially with the newer drives
(>100 or so GB) -
junk the drive and put a new one in.
In fact, based on what I've seen and heard I'd replace more of the drives once you
get the bad drive replaced
and the array rebuilt, especially if the drives are from the same batch. Marginal drives
can and do
go out during rebuilds. Of course, if the data on the array isn't critical then this
becomes much more flexible-
but if it is then you don't want to have your second drive go at some point in the
near future, then the third give
up while you're attempting to rebuild the array after the second disk's failure.
Any reason you're recovering data from the drive rather than rebuilding? I'd go
along with Eric here - if
you rebuild then you're starting with a known-good array, rather than potentially
having errors in your
replacement drive.
This isn't just anecdotes, either - I've seen numbers. That's one reason
I've gone to mirrors on my few critical
installations.
8 Oct
8 Oct
7:48 a.m.
On Sun, Oct 07, 2012 at 11:53:14PM -0400, Scott Quinn wrote:
In fact, based on what I've seen and heard I'd
replace more of the drives once you get the bad drive replaced
and the array rebuilt, especially if the drives are from the same batch. Marginal drives
can and do
go out during rebuilds. Of course, if the data on the array isn't critical then this
becomes much more flexible-
but if it is then you don't want to have your second drive go at some point in the
near future ...
RAID != backup
Not enough people remember this.
If a second disk dies in *my* RAID, it's annoying, but I don't lose
data. Well, I lose anything since the last backup, but given that those
happen every 24 hours, it's not *that* much data and I can reconstruct
it pretty easily.
--
David Cantrell | London Perl Mongers Deputy Chief Heretic
Vegetarian: n: a person who, due to poor lifestyle choices,
is more likely to get arse cancer than a normal person
8:09 a.m.
On Mon, Oct 08, 2012 at 03:48:41PM +0100, David Cantrell wrote:
On Sun, Oct 07, 2012 at 11:53:14PM -0400, Scott Quinn
wrote:
> In fact, based on what I've seen and heard I'd replace more of the drives
once you get the bad drive replaced
...
RAID != backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
Not enough people remember this.
Sad isn't it.
--
David Cantrell | London Perl Mongers Deputy Chief Heretic
Vegetarian: n: a person who, due to poor lifestyle choices,
is more likely to get arse cancer than a normal person
Diane
--
- db at FreeBSD.org db at db.net http://www.db.net/~db
Nowadays tar can compress using yesterdays latest technologies!
8:51 a.m.
On 8 October 2012 16:43, Dave McGuire <mcguire at neurotica.com> wrote:
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
Yeah, I've had an entire SAN die... and take all the disks with it.
--
Yoyodyne Propulsion Systems: "The Future Begins Tomorrow"
Visit us at: http://www.yoyodyne-propulsion.net
--------
"I'd like to hold off judgement on a thing like that, sir, until all
the facts are in. "
General "Buck" Turgidson
RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this. 
9:01 a.m.
On 10/08/2012 11:51 AM, John Many Jars wrote:
RAID != backup != offsite backup!
Yeah, I've had an entire SAN die... and take all the disks with it.
Eeeew. That sounds like a very long night!
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
RAID != backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this. 
8:56 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 08/10/12 11:43 AM, Dave McGuire wrote:
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
I consider RAID as an availability measure - but how many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
--T
RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
9:01 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/08/2012 11:56 AM, Toby Thain wrote:
RAID != backup != offsite backup!
I consider RAID as an availability measure - but how many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
I do. But then some people think I'm crazy for running grownup
equipment with grownup procedures "at home".
Granted "at home" in this case means a 14,000 square foot building
where a few dozen web sites and 100+ email accounts and lists are
hosted, as well as my "put dinner on the table" work, but still...if
it's worth doing, it's worth doing right.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
RAID != backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this. 
9:20 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Oct 8, 2012, at 12:01 PM, Dave McGuire wrote:
On 10/08/2012 11:56 AM, Toby Thain wrote:
I consider RAID as an availability measure - but how many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
I do. But then some people think I'm crazy for running grownup
equipment with grownup procedures "at home".
Granted "at home" in this case means a 14,000 square foot building
where a few dozen web sites and 100+ email accounts and lists are
hosted, as well as my "put dinner on the table" work, but still...if
it's worth doing, it's worth doing right.
Absolutely. My wife thinks I'm nuts for putting in the time and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
It was totally worth figuring out how to bend a 10-foot length of 2"
PVC conduit to go into the ceiling of our 6-foot basement, though (fun
fact: 2" PVC conduit can bend nearly 90 degrees if you work hard enough
at it). I'm not sorry at all, though; all our wiring goes neatly into
the basement where the big switch is, and I have full gigabit wiring to
my office on the second floor (and OM-3 fiber for when 10 gig is more
within my range).
I mean, I work from home, so there's that. She does see the ultimate
value in having good equipment, at least. :-)
- Dave
> RAID != backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
RAID != backup != offsite backup! 
9:38 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
Absolutely. My wife thinks I'm nuts for putting
in the time and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
It was totally worth figuring out how to bend a 10-foot length of 2"
PVC conduit to go into the ceiling of our 6-foot basement, though (fun
fact: 2" PVC conduit can bend nearly 90 degrees if you work hard enough
at it). I'm not sorry at all, though; all our wiring goes neatly into
the basement where the big switch is, and I have full gigabit wiring to
my office on the second floor (and OM-3 fiber for when 10 gig is more
within my range).
Wow. I just have my T1 run, which is 60' of two lengths of UV grade Cat5e
(one pair in one jacket, the other pair in the other jacket, with hand-wired
jacks) through external flex ducts, and then three main backbone lines
(10Mbit, 100Mbit and gig) between the server room and the office, wired into
the attic. I split up the traffic so that I could segregate them by speed.
The rest of the house uses relatively low-speed power line networking; I don't
trust wireless.
The T1 modem and the backbone switches and hubs sit in the server room,
plus the exit node for the powerline Ethernet.
I remember drilling the entry hole for the T1 lines and scaring a paper wasp
nest. As a pasty white nerd, I don't think I've ever run that fast.
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser at floodgap.com
-- I'm too old to use emacs. -- Rod MacDonald ---------------------------------
9:57 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Mon, Oct 8, 2012 at 12:38 PM, Cameron Kaiser <spectre at floodgap.com> wrote:
> Absolutely. My wife thinks I'm nuts for
putting in the time and effort
> to figure out how to run a bundle of Ethernet and fiber cables through
> the walls of our nearly 100-year-old house ("Don't we have wireless?").
I ran "interduct" (http://www.innerduct.com/products/corrugated.php) from my
basement to my back bedroom about 15 years ago and stuffed it with some
Cat5, some multi-mode fiber, two 25-pair Cat3 telco bundles (phone and
dumb terminals), and one 10Base2 for old stuff. There's still some room
in there for a few more bits.
> It was totally worth figuring out how to bend a
10-foot length of 2"
> PVC conduit to go into the ceiling of our 6-foot basement, though (fun
> fact: 2" PVC conduit can bend nearly 90 degrees if you work hard enough
> at it).
They now have a nice PVC heater/bender that I saw reviewed in Make
Magazine recently. You stuff the PVC in a corrugated tube, plug it in
and wait for the plastic to become soft and nearly limp. I presume one
removes the hot PVC and bends it to shape before it cools (the review
didn't make it clear if you are meant to remove power, then bend, then
cool, then remove, or some other order of operations). I don't know if
they make a 2" version, but the smaller version seemed to be quite
handy for trivial shaping.
-ethan
10:17 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Oct 8, 2012, at 12:57 PM, Ethan Dicks wrote:
They now have a nice PVC heater/bender that I saw
reviewed in Make
Magazine recently. You stuff the PVC in a corrugated tube, plug it in
and wait for the plastic to become soft and nearly limp. I presume one
removes the hot PVC and bends it to shape before it cools (the review
didn't make it clear if you are meant to remove power, then bend, then
cool, then remove, or some other order of operations). I don't know if
they make a 2" version, but the smaller version seemed to be quite
handy for trivial shaping.
Well, I wanted it to remain straight in the end; it just needed to
go straight up through the walls. The problem was one of simple
geometry, specifically inserting a 10-foot pipe vertically into the
wall-ceiling junction of a 6-foot basement. I didn't particularly
want to cut it, and the pipe resumed its previous shape fairly
quickly once I got it into the wall.
I was just lucky enough that the folks who had rehabbed the house
back in the '80s and added an HVAC system had left extra space
around the old chimney when they put in the duct work (and also
lucky that said space intersected my office). I ended up with
a built-in channel running straight from the basement to the
top floor for free; all I needed was to put a pipe in there so
the wires had some guidance going down.
- Dave
4:11 p.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Monday, October 8, 2012, Ethan Dicks <ethan.dicks at gmail.com> wrote:
On Mon, Oct 8, 2012 at 12:38 PM, Cameron Kaiser
<spectre at floodgap.com>
wrote:
They now have a nice PVC heater/bender that I saw
reviewed in Make
Magazine recently. You stuff the PVC in a corrugated tube, plug it in
and wait for the plastic to become soft and nearly limp. I presume one
removes the hot PVC and bends it to shape before it cools (the review
didn't make it clear if you are meant to remove power, then bend, then
cool, then remove, or some other order of operations). I don't know if
they make a 2" version, but the smaller version seemed to be quite
handy for trivial shaping.
I've used a propane torch from a distance back in the day, but i believe a
heatgun would work as well
10:03 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Oct 8, 2012, at 12:38 PM, Cameron Kaiser wrote:
I don't have problems with wireless per se, but it has its uses. When
we lived in a shorter house (which was rented, so I couldn't really
perform modifications if I wanted to), I ran most of my vintage equipment
on a LAN with a wireless bridge. Apple's Airport base stations actually
do a very nice job of transparently bridging Ethernet and wireless for
all purposes I've used them for so far.
However, our current house is too tall and too dense for the wireless
signal to be 100% reliable, and I've started doing enough high-bandwidth
networking stuff for work that I want to be able to segment off the
network into VLANs. Apparently throwing nearly 100 Mbps of multicast
traffic over a LAN makes the slower computers (which are innocent
bystanders) a little unhappy. :-\
As far as the fiber, OM-3 patch cables are only about a dollar a meter
at Monoprice, so I figured I might as well run three up to my office
just in case. I have a 12-port Keystone plate up there with 9 Cat5e
jacks and 3 fiber jacks.
I have little to no use for a T1 around here, since I'm not running a
server out of my house (I run one out of my parents' house, but it's
nothing particularly mission-critical, so they just have business
FiOS). Regarding a recent discussion, it's a dual-CPU Opteron with
12 GB RAM, which was roughly equivalent to the 8-core Xeon with 96 GB
RAM when it was built (around 2005ish?). It gets something like 100
hits a week, but I've found it invaluable to have a machine with a
reasonable amount of horsepower sitting around so I can run offline
computations (and now Minecraft servers).
- Dave
Absolutely.
My wife thinks I'm nuts for putting in the time and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
It was totally worth figuring out how to bend a 10-foot length of 2"
PVC conduit to go into the ceiling of our 6-foot basement, though (fun
fact: 2" PVC conduit can bend nearly 90 degrees if you work hard enough
at it). I'm not sorry at all, though; all our wiring goes neatly into
the basement where the big switch is, and I have full gigabit wiring to
my office on the second floor (and OM-3 fiber for when 10 gig is more
within my range).
Wow. I just have my T1 run, which is 60' of two lengths of UV grade Cat5e
(one pair in one jacket, the other pair in the other jacket, with hand-wired
jacks) through external flex ducts, and then three main backbone lines
(10Mbit, 100Mbit and gig) between the server room and the office, wired into
the attic. I split up the traffic so that I could segregate them by speed.
The rest of the house uses relatively low-speed power line networking; I don't
trust wireless.
The T1 modem and the backbone switches and hubs sit in the server room,
plus the exit node for the powerline Ethernet.
I remember drilling the entry hole for the T1 lines and scaring a paper wasp
nest. As a pasty white nerd, I don't think I've ever run that fast.
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser at floodgap.com
-- I'm too old to use emacs. -- Rod MacDonald ---------------------------------
11:22 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/08/2012 12:20 PM, David Riley wrote:
Absolutely. My wife thinks I'm nuts for putting in the time and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
She's obviously never had to move big files over a wireless network. ;)
I consider RAID as an availability measure - but how
many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
I do. But then some people think I'm crazy for running grownup
equipment with grownup procedures "at home".
Granted "at home" in this case means a 14,000 square foot building
where a few dozen web sites and 100+ email accounts and lists are
hosted, as well as my "put dinner on the table" work, but still...if
it's worth doing, it's worth doing right. It was totally worth figuring out how to bend a
10-foot length of 2"
PVC conduit to go into the ceiling of our 6-foot basement, though (fun
fact: 2" PVC conduit can bend nearly 90 degrees if you work hard enough
at it). I'm not sorry at all, though; all our wiring goes neatly into
the basement where the big switch is, and I have full gigabit wiring to
my office on the second floor (and OM-3 fiber for when 10 gig is more
within my range).
I mean, I work from home, so there's that. She does see the ultimate
value in having good equipment, at least. :-)
Sounds good to me!
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
3:48 p.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/08/2012 11:20 AM, David Riley wrote:
My wife thinks I'm nuts for putting in the time
and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
Just tell her that the alternative was to run it up the outside walls :-)
3:51 p.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Mon, Oct 8, 2012 at 6:48 PM, Jules Richardson
<jules.richardson99 at gmail.com> wrote:
On 10/08/2012 11:20 AM, David Riley wrote:
I'm lucky I have a passage along the sewer pipe that runs from the
basement to the attic. Do you have balloon framing and/or firebreaks
in the wall?
--
-Jon
Jonathan Katz, Indianapolis, IN.
My wife thinks I'm nuts for putting in the
time and effort
to figure out how to run a bundle of Ethernet and fiber cables through
the walls of our nearly 100-year-old house ("Don't we have wireless?").
Just tell her that the alternative was to run it up the outside walls :-)
9:33 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
I consider
RAID as an availability measure - but how many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
I do. But then some people think I'm crazy for running grownup
equipment with grownup procedures "at home". 
3:07 p.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Mon, 8 Oct 2012, Cameron Kaiser wrote:
So do I. I have spare disks for the two small arrays I run, and I even have
an entire new controller and backplane in case the array itself goes toes up.
I'll be getting a full body double for the POWER6 as soon as prices drop
enough.
Amen. Just in the process of building the new server. Picked up a Dell
PowerEdge 2950 and have been putting together one extra of everything off
of eBay. These are all coming off lease now and parts are quite
inexpensive.
I've been burned before with a 3ware SATA RAID controller dropping dead
(those things run HOT), so a backup PERC/5i was the first thing I went
after. Planning a Raid-5 array of WDC RE4 500GB drives with one hot-spare
and an extra drive on the supply shelf.
Even with the current server I backup to an LTO-2 drive on a regular
basis. One set of tapes is always in my desk at the office - five miles
away :-).
The new box has redundant power supplies and I have a third one waiting if
problems occur. The old server ate one high-end ATX supply per year on
the average - and these were NOT cheapies. Seems like all consumer
supplies are crap nowadays.
Steve
--
I consider RAID as an availability measure - but how
many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :)
I do. But then some people think I'm crazy for running grownup
equipment with grownup procedures "at home".
9:20 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Mon, Oct 8, 2012 at 11:56 AM, Toby Thain <toby at telegraphics.com.au> wrote:
The IceCube datacenter at Pole keeps dozens of spare disks on hand for
replacements, but there, you can't just phone someone up in June and
get a disk shipped in overnight... well... it'll be there half-past
sunrise, but that means "November" in this case.
One advantage of a remote site is that you *know* you can't get parts
in when you want them so you have to plan in advance and maintain a
spares pile. No other place I've worked kept spare hardware around
like that - when it dies, call it in and they send one (and you hope
that a second disk doesn't die before the first arrives). Had that
happen exactly once - disk 1 dies, call for replacement, 5 hours
later, disk 2 dies, replace 2 disks in moments then spend days
rebuilding the array from what backups there are.
-ethan
RAID !=
backup != offsite backup!
I consider RAID as an availability measure - but how many installations have
you seen that have spare disks on-hand and procedures to monitor and replace
them? (I am sure the serious ones do. Elsewhere, every cowboy sysadmin
configures RAID on day 1... but where are the disks? :)
10:33 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 08/10/12 12:20 PM, Ethan Dicks wrote:
On Mon, Oct 8, 2012 at 11:56 AM, Toby Thain<toby at
telegraphics.com.au> wrote:
The IceCube datacenter at Pole keeps dozens of spare disks on hand for
replacements, but there, you can't just phone someone up in June and
get a disk shipped in overnight... well... it'll be there half-past
sunrise, but that means "November" in this case.
One advantage of a remote site is that you *know* you can't get parts
in when you want them so you have to plan in advance and maintain a
spares pile. No other place I've worked kept spare hardware around
When ex-employer put in an Xserve G4 + Xserve RAID (2003?) we had spares
of everything. It was a $40,000 system at that time (1.1TB online).
It's not just time-to-order; suitable disks often get EOL'd before the
originals fail.
I just find it amusing that the cargo cult practice is to configure
servers with RAID without thinking it through. (Or realising that RAID
isn't state of the art.) :) (Obviously this doesn't apply to actual
professional outfits.)
--T
RAID
!= backup != offsite backup!
I consider RAID as an availability measure - but how many installations have
you seen that have spare disks on-hand and procedures to monitor and replace
them? (I am sure the serious ones do. Elsewhere, every cowboy sysadmin
configures RAID on day 1... but where are the disks? :) like that - when it dies, call it in and they send one
(and you hope
that a second disk doesn't die before the first arrives). Had that
happen exactly once - disk 1 dies, call for replacement, 5 hours
later, disk 2 dies, replace 2 disks in moments then spend days
rebuilding the array from what backups there are.
-ethan
9:55 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/8/12 10:56 AM, Toby Thain wrote:
On 08/10/12 11:43 AM, Dave McGuire wrote:
raises hand... I have spares as well as scripts that send me
activity, volume capacity/usage, & SMART status daily.
My bosses moan every time I order disks - why do we spend double on
storage for EVERY server???? (ZFS mirrors)
In a smallish storage environment, I've had 4 failed server disks in
3 years here, with no data loss and a total of 1 hour downtime. (ZFS,
for some reason, threw a kernel panic over one of the failed drives).
Also, I totally agree with whoever mentioned preemptively replacing
disks. At three years of 24/7 service, they get replaced. Drives are
cheap, my time and my users' data are not.
Doc
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
I consider RAID as an availability measure - but how many installations
have you seen that have spare disks on-hand and procedures to monitor
and replace them? (I am sure the serious ones do. Elsewhere, every
cowboy sysadmin configures RAID on day 1... but where are the disks? :) RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
10:27 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/8/12 12:04 PM, Al Kossow wrote:
On 10/8/12 9:55 AM, Doc Shipley wrote:
Mmmm.... about 150
Doc
I've had 4 failed server disks in 3 years
here, with no data loss and
a total of 1 hour downtime.
out of how many drives? 
10:53 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On 10/8/12 10:27 AM, Doc Shipley wrote:
On 10/8/12 12:04 PM, Al Kossow wrote:
Interesting data point. Drive count is about the same as what we're building
out right now
http://www.computerhistory.org/atchm/crafting-a-digital-repository/
On 10/8/12 9:55 AM, Doc Shipley wrote:
Mmmm.... about 150
I've had 4 failed server disks in 3 years
here, with no data loss and
a total of 1 hour downtime.
out of how many drives? 
9 Oct
9 Oct
4:56 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
Al Kossow <aek at bitsavers.org> wrote:
Interesting data point. Drive count is about the same
as what we're building
out right now
http://www.computerhistory.org/atchm/crafting-a-digital-repository/
Interesting read! Any ETA on the article on _why_ you chose Archivematica?
re,
sander
8 Oct
8 Oct
10:42 a.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
On Mon, Oct 08, 2012 at 11:55:10AM -0500, Doc Shipley wrote:
On 10/8/12 10:56 AM, Toby Thain wrote:
> On 08/10/12 11:43 AM, Dave McGuire wrote:
>> On 10/08/2012 11:09 AM, Diane Bruce wrote:
>>>> RAID != backup
...
Also, I totally agree with whoever mentioned preemptively replacing
disks. At three years of 24/7 service, they get replaced. Drives are
cheap, my time and my users' data are not.
Ugh. yes to this. in spades.
Doc
Diane
--
- db at FreeBSD.org db at db.net http://www.db.net/~db
Nowadays tar can compress using yesterdays latest technologies!
4:13 p.m.
New subject: Redundant Array of Inexpensive Sysadmins - Re: disk flaws, classic vs. modern
> It does
not help if you have a fire. A few
businesses have been lost due
>> to not remembering this.
>
>? ? RAID != backup != offsite backup! have you seen that have spare disks on-hand and
procedures
to monitor
and replace them? (I am sure the serious ones do. Elsewhere,
every
cowboy sysadmin configures RAID on day 1... but where are
the disks? :)
In the case of the products we build, the disks are
spares marked as such in the storage array. If a
drive dies in a redundant array (RAIDs 1, 5, 6, or
10), a spare is automatically pulled in and rebuilt.
BLS
9:40 a.m.
On 8 October 2012 16:43, Dave McGuire <mcguire at neurotica.com> wrote:
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
I setup a little HP MicroServer (very cute little box) with 5x2TB
disks in a RAID5, and hit just this dilemma. Solution was to buy a
second similarly equipped and use as the offsite backup.
Of course you then have to worry about the security model of someone
breaking into either your main box or offsite backup box (depending on
the direction of your security gradient....)
RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
10:37 a.m.
On Mon, Oct 08, 2012 at 11:43:31AM -0400, Dave McGuire wrote:
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
We are agreeing. It makes little sense to do a backup and leave the backups
in the same place the computer system can get burned up. ;-)
RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
--
Dave McGuire, AK4HZ
New Kensington, PA
Diane Bruce VA3DB
--
- db at FreeBSD.org db at db.net http://www.db.net/~db
Nowadays tar can compress using yesterdays latest technologies!
3:47 p.m.
> RAID != backup != offsite backup!
Entirely.
We are agreeing. It makes little sense to do a backup
and leave the
backups in the same place the computer system can get burned up. ;-)
That depends on why you're doing backups, the value of what you're
protecting, and the cost of offsite storage. In my case, that means no
RAID, but live mirroring to the backup machine, with live mirroring
from there to a removable disk which I swap weekly with
an identical
disk on my desk at work (a fifteen-minute bicycle ride away). I've
had
two? three? drives fail since I started doing that, and all the bits
have been safe each time.
But then, I've specifically decided to not keep backups for historical
archives or accidental deletion purposes (those being the other two
reasons I've been able to come up with to keep backups). Ocasionally
I've been sorry, but so far the cost of those incidents has been well
below the estimated cost of keeping backups suitable for handling them.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
10 Oct
10 Oct
6:58 p.m.
On 10/8/12 8:43 AM, "Dave McGuire" <mcguire at neurotica.com> wrote:
On 10/08/2012 11:09 AM, Diane Bruce wrote:
RAID != backup != offsite backup!
And management questioned when I designed the digital image archiving system
for a museum when I had the assets deposited to a farm of servers (each with
a RAID array), then copied to a Sun server with a large raid array and a
tape backup library that was running software that made the tapes look like
part of the RAID array, and pulling a backup of new data onto specially ID'd
tapes in the library that were spit out of the library once full and taken
to an offsite temp/humidity controlled storage facility. One of the reqs
for the system was that if a meteor hit the facility with the servers we
could after getting duplicate hardware installed in a new facility we could
restore from the tapes and be back online within 7 days
RAID !=
backup
It does not help if you have a fire. A few businesses have been lost due
to not remembering this.
4458
days inactive
4461
days old
31 comments
20 participants
participants (20)
-
abs＠absd.org -
aek＠bitsavers.org -
blstuart＠bellsouth.net -
david＠cantrell.org.uk -
db＠db.net -
doc＠vaxen.net -
ethan.dicks＠gmail.com -
fraveydank＠gmail.com -
geoffr＠zipcon.net -
jlobocki＠gmail.com -
john＠yoyodyne-propulsion.net -
jon＠jonworld.com -
jules.richardson99＠gmail.com -
mcguire＠neurotica.com -
mouse＠Rodents-Montreal.ORG -
reiche＠ls-al.eu -
saquinn624＠aol.com -
snhirsch＠gmail.com -
spectre＠floodgap.com -
toby＠telegraphics.com.au