Backup!

Just had a drive fail here on my PC. Drive has lots of important stuff
on it like irreplaceable photos, etc.

The good news is:

1) drive is under warranty

2) Drive was part of a RAID 1 array

Bad news is:

I have not backed up in .... forever (I know , SLAP me! This was VERY
stupid, and I only got saved this time because I have a RAID. Next time,
if it is a major failure or virus / worm I may not be so lucky without a
backup.)


So, lets ALL learn from this shall we?

Tomorrow I go out and buy an external 160 GB drive and back up
EVERYTHING using something like NORTON GHOST, and return this dead
MAXTOR SATA drive (which by the way is less than 8 months old) and after
a complete backup to the external drive, I get to rebuild my raid
array.... somehow.

Drive and system has been performing flawlessly since I built it. No
warning whatsoever that a drive was flakey. Still not sure if this is a
one-time glitch but until I back up EVERYTHING I will take no chances.

So friends, any bets whether the other identical MAXTOR drive (with a
serial number that is only 1 digit off from the failed drive) is going
to go 'poof' soon? (Not too soon I hope!)

God I hope I dont have to re-install the OS, (beg microsoft for
forgiveness etc.) install patches, load backup software and then restore
everything from the external drive!

At first glance it appears that a RAID 1 array (with no spare drives by
definition) is not easily rebuilt when a drive fails.

I am going to have to call my buddy Dave (the one with too much software
and hardware knowledge for any person, AND a digital photography habit)
and ask him how to handle this situation the best way. Dave is a great
guy. Dave likes wine. I will give him wine for saving my A$$ Nice
Dave. PLEASE!!!!!!

Steve

steve wrote:



At first glance it appears that a RAID 1 array (with no spare drives by
definition) is not easily rebuilt when a drive fails.


RAID 1 is simple mirroring. If the controller can't simply duplicate
the still working drive to the new one after it is installed then the
controller is seriously lacking.

Point taken about backing up.

Greg

--
Destroy your safe and happy lives
Before it is too late
The battles we fought were long and hard
Just not to be consumed by rock'n'roll

"steve" wrote

So friends, any bets whether the other identical MAXTOR drive (with a
serial number that is only 1 digit off from the failed drive) is going to go
'poof' soon? (Not too soon I hope!)

Can happen. I had a Seagate 15.3 Gb drive fail mechanically, and it was
swapped for an identical model by the local agent. It too failed within a
few months. A bigger Seagate bought to replace it is working perfectly.

I guess there are electronic lemons around...

BTW, I use 'Second Copy' for backing up my data daily - see www.centered.com

Paul

Agreed.

I have read the minimal manual that Intel provides for the ICH5R RAID
controller in this 875P motherboard chipset. There appears to be ZERO
support from Intel other than the manual. I suppose this is typical for
an OEM product. Comes with the territory.

From reading the manual for the RAID chipset it appears that I can
re-create the RAID 1 ARRAY as follows:

1. Backup everything (O/S, and Data, all partitions) to an additional
bootable hard drive. In other words make a bootable image of everything
on the RAID. (I sure hope Ghost can do this, and I believe it probably can.)

2. Delete the raid volume (This step destroys all data including the O/S
on the working RAID drive)

3. Boot to the backup drive, and launch the Intel data migration tool.

4. Create a new RAID 1 volume using the tool, which will migrate
everything from the backup drive to the new RAID volume.

What isnt clear is how and when (or even if I have to) partition and
format the new 'virgin' SATA drive in this process.

Hold me , I'm scared

Steve


G.T. wrote:

steve wrote:



At first glance it appears that a RAID 1 array (with no spare drives
by definition) is not easily rebuilt when a drive fails.


RAID 1 is simple mirroring. If the controller can't simply duplicate
the still working drive to the new one after it is installed then the
controller is seriously lacking.

Point taken about backing up.

Greg

"steve" skrev i melding
...
Agreed.

I have read the minimal manual that Intel provides for the ICH5R RAID
controller in this 875P motherboard chipset. There appears to be ZERO
support from Intel other than the manual. I suppose this is typical for
an OEM product. Comes with the territory.

From reading the manual for the RAID chipset it appears that I can
re-create the RAID 1 ARRAY as follows:

1. Backup everything (O/S, and Data, all partitions) to an additional
bootable hard drive. In other words make a bootable image of everything
on the RAID. (I sure hope Ghost can do this, and I believe it probably
can.)

2. Delete the raid volume (This step destroys all data including the O/S
on the working RAID drive)

3. Boot to the backup drive, and launch the Intel data migration tool.

4. Create a new RAID 1 volume using the tool, which will migrate
everything from the backup drive to the new RAID volume.


That sounds like the thing to do if you were migrating from non-RAID to
RAID.

You already had a RAID 1 array, and should only need to plug in the new disk
and let the ICH5R do its magic by copying everything from the old
functioning drive to the new one.

If you have installed the "Intel Application Accelerator RAID Edition", a
message wil pop up from the lower right corner of the screen, showing how
things are progressing. Synchronizing big disks may take several hours.

jpm

You might want to take a look at bootitng.com for their
partitioning/imaging/multi-os utility. It's very reasonably priced, you can
evaluate fully before you register and there is an excellent ng for
terabyteunlimited. I have no personal interest in the outfit apart from a
number of years happy useage of the product. Getting a usb hd is an
excellent idea these days although I haven't taken my own advice on this
yet.
Dave Cohen
"Paul Bartram" p.bartram AT OR NEAR optusnet.com wrote in message
u...

"steve" wrote

So friends, any bets whether the other identical MAXTOR drive (with a
serial number that is only 1 digit off from the failed drive) is going to
go
'poof' soon? (Not too soon I hope!)

Can happen. I had a Seagate 15.3 Gb drive fail mechanically, and it was
swapped for an identical model by the local agent. It too failed within a
few months. A bigger Seagate bought to replace it is working perfectly.

I guess there are electronic lemons around...

BTW, I use 'Second Copy' for backing up my data daily - see
www.centered.com

Paul

The manual for the Intel RAID software that came with the controller
built into the motherboard is strangely silent on how to recover from a
drive failure. I read every page last night. Carefully.

I will back everything up to an external drive and try what you have
suggested. If I am lucky the controller will make my life much easier.

If not, it is beginning to look like a pretty ugly process.

Do you have first hand experience performing a recovery of this nature
with this controller? I would greatly appreciate some tips if the answer
is yes!


Regards,

Steve


Jan-Petter Midtgård wrote:
"steve" skrev i melding
...

Agreed.

I have read the minimal manual that Intel provides for the ICH5R RAID
controller in this 875P motherboard chipset. There appears to be ZERO
support from Intel other than the manual. I suppose this is typical for
an OEM product. Comes with the territory.

From reading the manual for the RAID chipset it appears that I can
re-create the RAID 1 ARRAY as follows:

1. Backup everything (O/S, and Data, all partitions) to an additional
bootable hard drive. In other words make a bootable image of everything
on the RAID. (I sure hope Ghost can do this, and I believe it probably

can.)

2. Delete the raid volume (This step destroys all data including the O/S
on the working RAID drive)

3. Boot to the backup drive, and launch the Intel data migration tool.

4. Create a new RAID 1 volume using the tool, which will migrate
everything from the backup drive to the new RAID volume.



That sounds like the thing to do if you were migrating from non-RAID to
RAID.

You already had a RAID 1 array, and should only need to plug in the new disk
and let the ICH5R do its magic by copying everything from the old
functioning drive to the new one.

If you have installed the "Intel Application Accelerator RAID Edition", a
message wil pop up from the lower right corner of the screen, showing how
things are progressing. Synchronizing big disks may take several hours.

jpm

steve writes:

Just had a drive fail here on my PC. Drive has lots of important stuff on it
like irreplaceable photos, etc.

The good news is:

1) drive is under warranty

2) Drive was part of a RAID 1 array

Bad news is:

I have not backed up in .... forever (I know , SLAP me! This was VERY stupid,
and I only got saved this time because I have a RAID. Next time, if it is a
major failure or virus / worm I may not be so lucky without a backup.)


So, lets ALL learn from this shall we?

Tomorrow I go out and buy an external 160 GB drive and back up EVERYTHING
using something like NORTON GHOST, and return this dead MAXTOR SATA drive
(which by the way is less than 8 months old) and after a complete backup to
the external drive, I get to rebuild my raid array.... somehow.

DVD burners are getting cheap enough that you should also consider getting one,
and making two copies, and keeping one in an external location, such as at your
office, or in a safe deposit box. I just got one (~ $40-70 at Staples after
rebate) to replace the 50 or so CD-R backups I was at with 6 DVD-R's. Given
you have RAID-1 you should be able to have the OS rebuild the raid array (which
given it is a simple mirror, is just copying the disk). I prefer to use DVD-R
instead of DVD+R, since the older DVD readers will read the DVD. You can never
have enough backups. A Raid backup doesn't prevent problems like accidently
deleting the files or a virus.

At the moment, in addition to CD/DVD backups, I keep my pictures on 2
computers: the main storage server, a backup computer that mirrors the disk (I
have a subset of the pictures on the laptop and external web server, but not
enough to recover from a problem). When I load new pictures, I then update the
backup server (no nightly copies, etc.). The idea is I do the backups when I
think things are stable. If I accidently nuke the files on the main server, I
can just reload from the backup (which happened when I was experimenting with
new software to upload the web server, and I did a get operation instead of a
put operation). With Linux it is fairly easy to do this with ssh and rsync,
and I have my master perl script offer to do it as part of the upload and
post-processing. I don't know what effort it would be on Microsoft OSes.

--
Michael Meissner
email:
http://www.the-meissners.org

Steve,

Just had a drive fail here on my PC. Drive has lots of important stuff
on it like irreplaceable photos, etc.
So, lets ALL learn from this shall we?

Been there, done that... Ha Ha.

May I recommend you use a removable media such as DVD-RW or a tape backup.
This allows you to make multiple backups and store at least one off site (I
use a safe deposit box) to protect against fire, theft, etc.

I might also suggest that you make regular "permanent" backups, maybe every
month, every 6 months, or whatever. Consider what happens if you keep using
the same media for backups. Your photos, financial records, or whatever may
be corrupted on your hard drive (glitch, drive failure, virus, etc.) and
you may not realize the files have been changed. You overwrite your
previous "good" backup with the now corrupted backup. Essentially, you are
left with a worthless backup. If you keep backups that you never overwrite,
you'll guard against this situation. Again, learned from personal
experience...

Anthony

I have tested it by stopping the computer, disconnecting one drive to
simulate failure, restarted the computer and copyed files from one directory
to another etc on the remainig disk, stopped, reconnected and started again.
The RAID controller responded as described earlier, and in about an hour the
two disks were syncronized and running like they did before the test.

jpm




"steve" skrev i melding
...
The manual for the Intel RAID software that came with the controller
built into the motherboard is strangely silent on how to recover from a
drive failure. I read every page last night. Carefully.

I will back everything up to an external drive and try what you have
suggested. If I am lucky the controller will make my life much easier.

If not, it is beginning to look like a pretty ugly process.

Do you have first hand experience performing a recovery of this nature
with this controller? I would greatly appreciate some tips if the answer
is yes!


Regards,

Steve


Jan-Petter Midtgård wrote:
"steve" skrev i melding
...

Agreed.

I have read the minimal manual that Intel provides for the ICH5R RAID
controller in this 875P motherboard chipset. There appears to be ZERO
support from Intel other than the manual. I suppose this is typical for
an OEM product. Comes with the territory.

From reading the manual for the RAID chipset it appears that I can
re-create the RAID 1 ARRAY as follows:

1. Backup everything (O/S, and Data, all partitions) to an additional
bootable hard drive. In other words make a bootable image of everything
on the RAID. (I sure hope Ghost can do this, and I believe it probably

can.)

2. Delete the raid volume (This step destroys all data including the O/S
on the working RAID drive)

3. Boot to the backup drive, and launch the Intel data migration tool.

4. Create a new RAID 1 volume using the tool, which will migrate
everything from the backup drive to the new RAID volume.



That sounds like the thing to do if you were migrating from non-RAID to
RAID.

You already had a RAID 1 array, and should only need to plug in the new
disk
and let the ICH5R do its magic by copying everything from the old
functioning drive to the new one.

If you have installed the "Intel Application Accelerator RAID Edition",
a
message wil pop up from the lower right corner of the screen, showing
how
things are progressing. Synchronizing big disks may take several hours.

jpm