R
Ryan Underwood
Is it just me, or does Maxtor suddenly have a huge stick up their ass with
respect to warranty RMA's? Specifically, having requested replacements
for ATA drives as recently as last year, and SCSI drives as recently as
a month ago, I can never remember them insisting on a diagnostic code to
process a warranty. This is a problem because they have no Linux version
of their diagnostic tool, so I would have to take the entire server down
to run the diagnostic on a single failed drive out of an array.
I tried to do the RMA on their website, but got hung up with the
diagnostic code requirement. Remembering that WDC used to require a
diagnostic code for web RMA's but didn't bother you if you did a phone
RMA, I called up. Well, the lackey I got insisted on the diagnostic code
giving no particular reason why, even when I explained the situation, so I
asked to speak to his manager. After 10 minutes on hold, the
conversation went mostly like this:
--------------------------------------
"Sir, we need a diagnostic code to process the RMA."
"Well, I just returned a drive not more than a month ago, and they didn't
hassle me about any diagnostic code."
"Well, that means someone didn't do their job."
.... "The drive went down out of an array for bad blocks. It needs to be
replaced."
"Well sir, if you run the diagnostic utility, it will fix those bad
blocks, and then you won't need to return the drive."
.... (having heard this line before) "Please explain to me how a software
utility will correct a media defect."
"Sir, be advised that media problems are not covered under warranty and
you will need to contact your media software vendor."
.... (incredulously) "No, that's not what I mean. I mean there is a
problem with the physical medium that the drive is storing information on.
I want you to explain to me how a software utility would "fix" a defect
rather than just mask an underlying problem.
"Sir, if you just run that utility, it will fix the problem."
"Even if it were able to do so, I would have to take the entire server
down to do that, which defeats the point of a hot spare, and causes me to
lose money."
"Well then you're going to have to do that. Your account is now flagged
for a bad blocks complaint, and any time you call back for any drive in
the future, we will have to insist on a diagnostic code before proceeding."
.... (thinking fast) "Well, the drive just died. I guess I won't be able
to run the diagnostic now."
"Ok sir, now if you would just verify your address and telephone number,
I'll have someone prepare the RMA for you and give you a call back."
----------------------------------------------------
Is this ridiculous or what? Not only the quick change of tune, but the
insistence on the diagnostic utility's ability to fix a problem. I have
never seen a drive be magically repaired by a manufacturer's utility.
Usually the problem just ends up coming back later, so you get to lose
whatever was stored on that block twice. I knew people who used IBM
75GXP drives in their servers, and when they eventually crapped out, they
would run DFT, it would "fix" the drive, and then they would go on their
merry way, only to have the drive crap out again a few weeks later.
Thank $DEITY for RAID-1! Of course, that doesn't help you if all the
drives in the set are full of holes.
My theory is that the "repair" in any manufacturer's utility simply writes
data to that block that exercises the magnetic field in that area, in an
attempt to shake out problems related to magnetic drift. But what's the
likelihood of normal magnetic drift being a problem on a drive less than 2
years old, compared to an actual media defect which causes the block
mapped at that location not to be able to hold onto its data?
Usually modern drives are doing block remapping behind the scenes too;
when a block has a correctable read error, the drive transparently moves
the data to a spare block and marks the original one bad. You can tell
when this has happened too, because reads that would previously have been
linear end up having intermittent seeks in the middle of the file, where
the drive seeks to the spare blocks(s) that now hold parts of that
file, and then back to where the rest of the file resides. So what would
the software utility magically accomplish that the drive's defect
management firmware would not?
The manager as well as the person that called back and actually wrote up
the RMA insisted that this policy has been in place for several years now.
I know that's either misinformed or a flat-out lie, because I had an
ATA drive replaced last year without any diagnostic, and the most recent
SCSI replacement was about a month ago, and they didn't want any
diagnostic, and the drive was even out of warranty when they replaced it.
I also fairly clearly remember submitting their web RMA form in the past
without needing a diagnostic code, so the "Diagnostic Code Required" step
in there is a new one on me.
This really wouldn't be a problem if they offered versions of their
diagnostic software that worked natively under Linux or whatever server OS
you are using. Asking to take your busy server down to run a diagnostic
(and then possibly again to install the new drive, without a hotplug
array) is just plain unreasonable in my opinion.
Thoughts? Am I just plain wrong and they have always required this, or
are they really tightening their belts at the expense of their customers'
patience and businesses?
Seagate is looking awful good these days.
respect to warranty RMA's? Specifically, having requested replacements
for ATA drives as recently as last year, and SCSI drives as recently as
a month ago, I can never remember them insisting on a diagnostic code to
process a warranty. This is a problem because they have no Linux version
of their diagnostic tool, so I would have to take the entire server down
to run the diagnostic on a single failed drive out of an array.
I tried to do the RMA on their website, but got hung up with the
diagnostic code requirement. Remembering that WDC used to require a
diagnostic code for web RMA's but didn't bother you if you did a phone
RMA, I called up. Well, the lackey I got insisted on the diagnostic code
giving no particular reason why, even when I explained the situation, so I
asked to speak to his manager. After 10 minutes on hold, the
conversation went mostly like this:
--------------------------------------
"Sir, we need a diagnostic code to process the RMA."
"Well, I just returned a drive not more than a month ago, and they didn't
hassle me about any diagnostic code."
"Well, that means someone didn't do their job."
.... "The drive went down out of an array for bad blocks. It needs to be
replaced."
"Well sir, if you run the diagnostic utility, it will fix those bad
blocks, and then you won't need to return the drive."
.... (having heard this line before) "Please explain to me how a software
utility will correct a media defect."
"Sir, be advised that media problems are not covered under warranty and
you will need to contact your media software vendor."
.... (incredulously) "No, that's not what I mean. I mean there is a
problem with the physical medium that the drive is storing information on.
I want you to explain to me how a software utility would "fix" a defect
rather than just mask an underlying problem.
"Sir, if you just run that utility, it will fix the problem."
"Even if it were able to do so, I would have to take the entire server
down to do that, which defeats the point of a hot spare, and causes me to
lose money."
"Well then you're going to have to do that. Your account is now flagged
for a bad blocks complaint, and any time you call back for any drive in
the future, we will have to insist on a diagnostic code before proceeding."
.... (thinking fast) "Well, the drive just died. I guess I won't be able
to run the diagnostic now."
"Ok sir, now if you would just verify your address and telephone number,
I'll have someone prepare the RMA for you and give you a call back."
----------------------------------------------------
Is this ridiculous or what? Not only the quick change of tune, but the
insistence on the diagnostic utility's ability to fix a problem. I have
never seen a drive be magically repaired by a manufacturer's utility.
Usually the problem just ends up coming back later, so you get to lose
whatever was stored on that block twice. I knew people who used IBM
75GXP drives in their servers, and when they eventually crapped out, they
would run DFT, it would "fix" the drive, and then they would go on their
merry way, only to have the drive crap out again a few weeks later.
Thank $DEITY for RAID-1! Of course, that doesn't help you if all the
drives in the set are full of holes.
My theory is that the "repair" in any manufacturer's utility simply writes
data to that block that exercises the magnetic field in that area, in an
attempt to shake out problems related to magnetic drift. But what's the
likelihood of normal magnetic drift being a problem on a drive less than 2
years old, compared to an actual media defect which causes the block
mapped at that location not to be able to hold onto its data?
Usually modern drives are doing block remapping behind the scenes too;
when a block has a correctable read error, the drive transparently moves
the data to a spare block and marks the original one bad. You can tell
when this has happened too, because reads that would previously have been
linear end up having intermittent seeks in the middle of the file, where
the drive seeks to the spare blocks(s) that now hold parts of that
file, and then back to where the rest of the file resides. So what would
the software utility magically accomplish that the drive's defect
management firmware would not?
The manager as well as the person that called back and actually wrote up
the RMA insisted that this policy has been in place for several years now.
I know that's either misinformed or a flat-out lie, because I had an
ATA drive replaced last year without any diagnostic, and the most recent
SCSI replacement was about a month ago, and they didn't want any
diagnostic, and the drive was even out of warranty when they replaced it.
I also fairly clearly remember submitting their web RMA form in the past
without needing a diagnostic code, so the "Diagnostic Code Required" step
in there is a new one on me.
This really wouldn't be a problem if they offered versions of their
diagnostic software that worked natively under Linux or whatever server OS
you are using. Asking to take your busy server down to run a diagnostic
(and then possibly again to install the new drive, without a hotplug
array) is just plain unreasonable in my opinion.
Thoughts? Am I just plain wrong and they have always required this, or
are they really tightening their belts at the expense of their customers'
patience and businesses?
Seagate is looking awful good these days.
