Discussion:
My Wednesday
(too old to reply)
T
2018-03-17 02:09:47 UTC
Permalink
Hi All,

Windows 7, sp1, x64

Wednesday was a fun day. Well, sort of…

A customer with RSTe RAID1 pair had one of the drives completely
fail. They were Intel SSDs. Quelle surprise! And yes they
were the ones Intel Tech support told me to use.

Now ordinarily, this is a pretty easy fix. You power down, remove
the replace the bad hard drive (Samsung this time), power back up
and use the Intel RST software utility to rebuild to a new drive.
Done it several times. It is usually boring.

NOT THIS TIME. The rebuild ran for about two minutes and them
the computer rebooted and Windows told me I could only boot into
Safe Mode. Okay, so I booted into safe mode, set up a
`chkdsk c: /f` and rebooted. This time Windows booted up okay.

Rise, later, repeat! AAAAAHHHHHH.

Now no drives register in the second slot. Hmmm. I knocked
off a power connector. Sheez.

Okay, so I rebooted, when into RSTe BIOS and broke the RAID
container. Now I am booting up as a single drive.

So, if the RSTe software did not work, I used the Samsung Data
Migration utility (doesn’t work on RAID pairs). I ran for about
6% and froze. What the hell???? Rebooted and tried again.
Same result. This is the FIRST time this utility has failed.
It is usually totally boring to run.

So, not to be defeated, I boot into Clone Zilla. I starts to clone
and Clone Zilla blows a cork on me. This time it tells me what
is wrong. Tons of bad sectors. Intel strikes again!!!!

Okay, so I rerun the clone and go into “Expert” mode and select
“rescue” (got to LOVE Clone Zilla). Rescue mode just skips over
the bad spots. This time I get a successful clone. I run gparted
from Fedora Xfce Live and expand the partition. And Windows likes
it too after a forced chkdsk. (Clone Zilla deliberately throws
the NTFS dirty flag).

So, I go into the BIOS RSTe and set up the RAID1 container. Ooops,
I forgot the stinking utility does not do the copy existing drive.
It should, but doesn’t get me started. Bad word. Bad word. Bad
word.

So, I had to clone everything over again. This time I use the
Windows RSTe utility to create the container.

And, FINALLY, everything is working again. I arrived at 15:39 and
left at 23:38.

That was my Wednesday. How was yours?

-T
VanguardLH
2018-03-17 02:50:01 UTC
Permalink
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
<snipped the personal diary entry>
Post by T
That was my Wednesday. How was yours?
This is Usenet. Not Facebook, Twitter, CafeMom, nor a logbook to store
your personal diary entries.

So does your *customer* paying you know about your shotgun repair?
T
2018-03-17 08:55:30 UTC
Permalink
Post by VanguardLH
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
<snipped the personal diary entry>
Post by T
That was my Wednesday. How was yours?
This is Usenet. Not Facebook, Twitter, CafeMom, nor a logbook to store
your personal diary entries.
So does your *customer* paying you know about your shotgun repair?
Are you having a bad day?
T
2018-03-17 09:02:59 UTC
Permalink
Post by T
Post by VanguardLH
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day.  Well, sort of…
<snipped the personal diary entry>
Post by T
That was my Wednesday.  How was yours?
This is Usenet. Not Facebook, Twitter, CafeMom, nor a logbook to store
your personal diary entries.
So does your *customer* paying you know about your shotgun repair?
Are you having a bad day?
Maybe worse than my Wednesday?

Today I am recoding a 6282 line Perl 5 program
into Perl 6. It is very slow going.

Perl 5's sub declarations are sh*t.
Char Jackson
2018-03-18 17:02:21 UTC
Permalink
Post by VanguardLH
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
<snipped the personal diary entry>
Post by T
That was my Wednesday. How was yours?
This is Usenet. Not Facebook, Twitter, CafeMom, nor a logbook to store
your personal diary entries.
So does your *customer* paying you know about your shotgun repair?
+1

IMHO, he shouldn't be billing anyone for that level of "service".
--
Char Jackson
B00ze
2018-03-17 03:42:54 UTC
Permalink
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
A customer with RSTe RAID1 pair had one of the drives completely
fail. They were Intel SSDs. Quelle surprise! And yes they
were the ones Intel Tech support told me to use.
Intel's are bad, they deliberatly go into DOA mode when they decide
there's too many bad blocks.
Post by T
Now ordinarily, this is a pretty easy fix. You power down, remove
the replace the bad hard drive (Samsung this time), power back up
and use the Intel RST software utility to rebuild to a new drive.
Done it several times. It is usually boring.
Doesn't it rebuild by itself? It did here on my RAID5 when I removed a
drive, booted into BIOS, shutdown, reconnected the drive and rebooted
into Windows. I never had to tell it to, it started to rebuild all by
itself.
Post by T
NOT THIS TIME. The rebuild ran for about two minutes and them
the computer rebooted and Windows told me I could only boot into
Safe Mode. Okay, so I booted into safe mode, set up a
`chkdsk c: /f` and rebooted. This time Windows booted up okay.
Okay, so I rebooted, when into RSTe BIOS and broke the RAID
container. Now I am booting up as a single drive.
You're lucky breaking the RAID left the drives bootable at all.
Post by T
So, if the RSTe software did not work, I used the Samsung Data
Migration utility (doesn’t work on RAID pairs). I ran for about
6% and froze. What the hell???? Rebooted and tried again.
Same result. This is the FIRST time this utility has failed.
It is usually totally boring to run.
So, not to be defeated, I boot into Clone Zilla. I starts to clone
and Clone Zilla blows a cork on me. This time it tells me what
is wrong. Tons of bad sectors. Intel strikes again!!!!
Never used CloneZilla, but I added it to my "Get This" list.
Post by T
Okay, so I rerun the clone and go into “Expert” mode and select
“rescue” (got to LOVE Clone Zilla). Rescue mode just skips over
the bad spots. This time I get a successful clone. I run gparted
from Fedora Xfce Live and expand the partition. And Windows likes
it too after a forced chkdsk. (Clone Zilla deliberately throws
the NTFS dirty flag).
So, I go into the BIOS RSTe and set up the RAID1 container. Ooops,
I forgot the stinking utility does not do the copy existing drive.
It should, but doesn’t get me started. Bad word. Bad word. Bad
word.
It's a BIOS extension, don't expect too much...
Post by T
So, I had to clone everything over again. This time I use the
Windows RSTe utility to create the container.
And, FINALLY, everything is working again. I arrived at 15:39 and
left at 23:38.
That was my Wednesday. How was yours?
Better than yours, but my Friday was just as bad as your Wednesday.

Regards,
--
! _\|/_ Sylvain / ***@hotmail.com
! (o o) Member:David-Suzuki-Fdn/EFF/Red+Cross/SPCA/Planetary-Society
oO-( )-Oo "Life," said Marvin, "Don't talk to ME about life!"
T
2018-03-17 08:57:11 UTC
Permalink
Post by B00ze
Now ordinarily, this is a pretty easy fix.  You power down, remove
the replace the bad hard drive (Samsung this time), power back up
and use the Intel RST software utility to rebuild to a new drive.
Done it several times.  It is usually boring.
Doesn't it rebuild by itself? It did here on my RAID5 when I removed a
drive, booted into BIOS, shutdown, reconnected the drive and rebooted
into Windows. I never had to tell it to, it started to rebuild all by
itself.
With RSTe and a new drive it does not recognize, you have to
press the "rebuild to another drive" button. It is pretty
simple.
Paul in Houston TX
2018-03-17 04:50:23 UTC
Permalink
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
Interesting post.
Thanks!
Paul
2018-03-17 07:40:48 UTC
Permalink
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day. Well, sort of…
A customer with RSTe RAID1 pair had one of the drives completely
fail. They were Intel SSDs.
I hope you're learning something from all this.

The very first error, is Intel drives with the
"fold up shop" behavior. If you were going to
RAID something, should have RAIDed an Intel and
a Sammy. Not two Intels. Not two *identical* Intels.

The possibility of correlated failures (both exceed
wear life in the same microsecond) should have entered
your mind. Each device in the RAID, does the same write
at the same time. Each wear counter wears down
at exactly the *same* rate. BOOM! BOOM! And
out go the lights.

You should not RAID two "fold up shop" drives
which are identical, because they will do the
identical math and fold up shop... together :-/

The Sammy probably doesn't fold up shop.

And the standard lessons have been taught here,
about SMART passthru. You mean to tell me, that in
addition to "degrade" and "fail", there's no
"SMART indicates imminent failure on Drive X"
message ? The array driver must be able to read
the SMART table, and pass a warning to the OS
in some way. If there is array management software,
a tray application, it should be installed and running.
(Some Promise installs, you could install the driver
and not bother with the tray. Driving blind.)

We already learned enough from the SIL3112 failure
modes, to trust soft (driver) RAID1 about as far as
it can be thrown.

*******

Did you compare the checksums on the unchanged
files in the most recent backup, against your clone
attempt ? You can use hashdeep to traverse a tree
and compute checksums.

Paul
T
2018-03-17 09:00:55 UTC
Permalink
Post by Paul
Post by T
Hi All,
Windows 7, sp1, x64
Wednesday was a fun day.  Well, sort of…
A customer with RSTe RAID1 pair had one of the drives completely
fail.  They were Intel SSDs.
I hope you're learning something from all this.
The very first error, is Intel drives with the
"fold up shop" behavior. If you were going to
RAID something, should have RAIDed an Intel and
a Sammy. Not two Intels. Not two *identical* Intels.
Intel SSD's are sh*t. I have taken it in the short
over them before. I will be glad when they are all gone.

I have said this before on these parts and some have argued
with me.
Post by Paul
Did you compare the checksums on the unchanged
files in the most recent backup, against your clone
attempt ? You can use hashdeep to traverse a tree
and compute checksums.
No. I just tried everything out afterwards and it
worked. The user is very studios and complains
a lot if anything is out of place. And she
was quite for two days.
Stan Brown
2018-03-17 11:44:36 UTC
Permalink
Post by Paul
I hope you're learning something from all this.
The very first error, is Intel drives with the
"fold up shop" behavior.
It drives me up the wall -- customers get a sequence of error
messages, and they tell me only about the last one. I don't know how
many times I've told someone, "When you get multiple error messages,
it's almost always the first one that gives the clue to what is
wrong."
--
Stan Brown, Oak Road Systems, Tompkins County, New York, USA
http://BrownMath.com/
http://OakRoadSystems.com/
Shikata ga nai...
T
2018-03-17 12:04:55 UTC
Permalink
Post by Stan Brown
Post by Paul
I hope you're learning something from all this.
The very first error, is Intel drives with the
"fold up shop" behavior.
It drives me up the wall -- customers get a sequence of error
messages, and they tell me only about the last one. I don't know how
many times I've told someone, "When you get multiple error messages,
it's almost always the first one that gives the clue to what is
wrong."
The first sign of trouble was when the second drive
in the pair because a brick. It showed as being there
but was unable to respond to anything.

The first drive stayed working, but was full of
sector errors: about 20 of them.

I bet the wear leveling had a joyful time of that!

Intel drive are sh*t. I wish I'd never got involved
with them in the first place. I only sell Samsung now
and have had ZERO issue with them. (It mesmerized
me that some argue about this with me!)

If you do any cloning, CloneZilla is a must in
your arsenal.

http://clonezilla.org/downloads.php

Expert mode, was a "rescue" switch that will clone
past errors. Often times, I am cloning to save
data from a dying drive.
Mr. Man-wai Chang
2018-03-17 11:50:28 UTC
Permalink
Post by T
....
And, FINALLY, everything is working again. I arrived at 15:39 and
left at 23:38.
That was my Wednesday. How was yours?
Did you got paid? :)
--
@~@ Remain silent! Drink, Blink, Stretch! Live long and prosper!!
/ v \ Simplicity is Beauty!
/( _ )\ May the Force and farces be with you!
^ ^ (x86_64 Ubuntu 9.10) Linux 2.6.39.3
不借貸! 不詐騙! 不賭錢! 不援交! 不打交! 不打劫! 不自殺! 不求神! 請考慮綜援
(CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa
T
2018-03-17 12:10:28 UTC
Permalink
Post by Mr. Man-wai Chang
....
And, FINALLY, everything is working again.  I arrived at 15:39 and
left at 23:38.
That was my Wednesday.  How was yours?
Did you got paid? :)
Not yet. I do my billing on Sundays usually.

This was a "mission critical" computer. A crash
would have been an absolute disaster for the company.
(Yes, it is backed up. The downtime would have killed
them.) That was shy the RAID1 in the first place.
I saved their asses.

I was glad I did not have to reinstall Windows and
all their apps. I would have been there all night!

Intel SSD's are sh*t.
Rene Lamontagne
2018-03-17 14:50:46 UTC
Permalink
 > ....
And, FINALLY, everything is working again.  I arrived at 15:39 and
left at 23:38.
That was my Wednesday.  How was yours?
Did you got paid? :)
Not yet.  I do my billing on Sundays usually.
This was a "mission critical" computer.  A crash
would have been an absolute disaster for the company.
(Yes, it is backed up.  The downtime would have killed
them.)  That was shy the RAID1 in the first place.
I saved their asses.
I was glad I did not have to reinstall Windows and
all their apps.  I would have been there all night!
Intel SSD's are sh*t.
Did you mention Intel drives are sh*t. :-)

Rene
Paul
2018-03-17 17:08:51 UTC
Permalink
Post by Rene Lamontagne
Post by T
Post by Mr. Man-wai Chang
Post by T
....
And, FINALLY, everything is working again. I arrived at 15:39 and
left at 23:38.
That was my Wednesday. How was yours?
Did you got paid? :)
Not yet. I do my billing on Sundays usually.
This was a "mission critical" computer. A crash
would have been an absolute disaster for the company.
(Yes, it is backed up. The downtime would have killed
them.) That was shy the RAID1 in the first place.
I saved their asses.
I was glad I did not have to reinstall Windows and
all their apps. I would have been there all night!
Intel SSD's are sh*t.
Did you mention Intel drives are sh*t. :-)
Rene
It's one particular policy they have, for device operation,
that makes an Intel SSD, a consumer-antagonistic device.

In a corporate environment, where fancy backup systems do incremental
backups on every disk in the company once a day, the Intel
policy doesn't matter. But for a home user who never makes
backups... the Intel approach is deadly.

When a device is near dead, it's OK for it to stop writing,
causing Windows to freak out. As long as it allows reading,
it allows a user to run a backup of the files. Then, once
a new SSD is installed, you can restore from backup. If the device
completely disappears without saying a word, well, that isn't
a very good policy. Lazy or unmotivated users, who have never
made a backup in their life, will learn a very sharp lesson
with their Intel SSD.

This is why, with SSDs, you really need to check each brand,
to see how they handle the "Intel issue".

The blue line here, the drive "pulled the trigger on itself"
when no bad sectors were present. The other three brands
continued to run,

https://techreport.com/review/27909/the-ssd-endurance-experiment-theyre-all-dead

That article covers a number of issues. It shows drives having
uncorrectable errors. It shows drives operating past their
calculated lifespan. And it mentions that once the lifespan
is surpassed, some notification was presented by the drive
that it was at the end-of-life conditions (wear limit).

You'll notice there's very little they could write about
the Intel drive, because it packed up before showing any
reallocations at all.

Paul
T
2018-03-18 00:46:43 UTC
Permalink
Post by Rene Lamontagne
Did you mention Intel drives are sh*t. :-)
Crap. My bad. It was just an oversight on my
part. I will do better next time!

:-)
Loading...