disk failure after PC was switched off

[Andy Fish]

Hi,

I can't remember the last time I switched my PC off - probably 6 months or
more ago, but I decided to leave it off over easter. When I turned it back
on again there were disk read errors and I ended up having to replace the
disk.

I assume that the shock of being shut down and restarted was what
precipitated the problem, and I have heard of this happening before on
server machines that never get shut down.

Should I be proactively shutting my PC down every so often to reduce the
chances of this happening, and if so, how often? or was I just really
unlucky?

Andy

 

[Jonny]

Hi,

I can't remember the last time I switched my PC off - probably 6 months or
more ago, but I decided to leave it off over easter. When I turned it back
on again there were disk read errors and I ended up having to replace the
disk.

I assume that the shock of being shut down and restarted was what
precipitated the problem, and I have heard of this happening before on
server machines that never get shut down.

Should I be proactively shutting my PC down every so often to reduce the
chances of this happening, and if so, how often? or was I just really
unlucky?

Andy

Dunno. Can just easily be said the PC on all the time was masking a looming
fault only seen at boot time.

 

[Arno Wagner]

Dunno. Can just easily be said the PC on all the time was masking a looming
fault only seen at boot time.

I agree. You should run a full smart self-test every two weeks or so
on an allways-on system. That way you will get early warning. You
should also log the smart status regularly, that way you have a better
chance of finding out what the actual problem was. It might just be that
the affected disk area had a sector that was slowly getting weak.
If it is read just once before it becomes completely unreadable
the disk will reallocate it. That is the second benefit of regular
surface scans (as a long smart self-test does).

From my experience the only component really at risk during a
rare power-cycle is the PSu, since it is under much more load
during the first seconds after power-on. All other components
are not that critical.

Arno

 

[Rod Speed]

I can't remember the last time I switched my PC off - probably 6 months
or more ago, but I decided to leave it off over easter.

Was that for lent ? |-(

When I turned it back on again there were disk read errors and I ended up
having to replace the disk.

I assume that the shock of being shut down and restarted was what
precipitated the problem,

Its much more likely that the drive was dying and either
the SMART check only happens at boot time or that the
section of the hard drive that was going bad only affected
boot ops so it was going bad and you didnt notice till a reboot.

and I have heard of this happening before on
server machines that never get shut down.

Yes, it can, for the same reason.

Should I be proactively shutting my PC down every so often to reduce the
chances of this happening,

Not in that sense, it doesnt reduce
anything, just gives you more warning.

and if so, how often? or was I just really unlucky?

Nope, just a bit unlucky that since you didnt reboot
much, you never noticed that the drive was dying.

One alternative is to use a better SMART app
that doesnt just check at boot time, or check
the SMART data occasionally, manually instead.

 

[Jonny]

I agree. You should run a full smart self-test every two weeks or so
on an allways-on system. That way you will get early warning. You
should also log the smart status regularly, that way you have a better
chance of finding out what the actual problem was. It might just be that
the affected disk area had a sector that was slowly getting weak.
If it is read just once before it becomes completely unreadable
the disk will reallocate it. That is the second benefit of regular
surface scans (as a long smart self-test does).

From my experience the only component really at risk during a
rare power-cycle is the PSu, since it is under much more load
during the first seconds after power-on. All other components
are not that critical.

Arno

No offense or counterpoint meant here.
So for usual folks that read/post at windowsxp.general group, more likely
they don't ever turn their PCs off (standby/hibernate). More than likely,
they have no access or inclination to use any software that checks hard
drive smart data in XP.
So, would these same people turning a PC on and off on daily basis where the
bios does the smart check or the hard drive(s) be a good idea for these?

 

[Arno Wagner]

No offense or counterpoint meant here. So for usual folks that
read/post at windowsxp.general group, more likely they don't ever
turn their PCs off (standby/hibernate). More than likely, they have
no access or inclination to use any software that checks hard drive
smart data in XP.

Well, reality does not care much what people's inclination is. These
people simply operate their machines at higher risk. But today many
people do. For example a recent test in the german computer magazine
c't found thet of 30 commercial windows backup programs a measly 3
were usable and reliable. The others had basic problems like silent
failures, no verify option or incomplete backups. It seems the average
windows user has no interest in keeping data safe or no clue that data
on a computer is permanently at risk. The horros stories we hear here
regularly (no backup, no SMART checks, no RAID and now the disk has
failed) are likely only the tip of the iceberg.

So, would these same people turning a PC on and off on daily basis
where the bios does the smart check or the hard drive(s) be a good
idea for these?

Unfortunately the BIOS does only check the SMART status, it does not
do a long SMART self-test. That means that rarely used files are still
at risk. This will however make continued bootability and early
detection of systematic problems (i.e. those not limited to specific
sectors) more likely. And it will make PC faliure more likely to
occure in ordinary, as opposed to special, circumstances. Could be a
good thing.

Arno

 

[DRE]

In comp.sys.ibm.pc.hardware.storage Jonny



Well, reality does not care much what people's inclination is. These
people simply operate their machines at higher risk. But today many
people do. For example a recent test in the german computer magazine
c't found thet of 30 commercial windows backup programs a measly 3
were usable and reliable.

Thats radically overstating it.

The others had basic problems like silent failures,

That doesnt make them unusable, they just
need more supervision and log checking.

no verify option or incomplete backups. It seems the average
windows user has no interest in keeping data safe or no clue
that data on a computer is permanently at risk.

Thats silly too. Plenty do have backups.

The horros stories we hear here regularly (no backup,
no SMART checks, no RAID and now the disk has
failed) are likely only the tip of the iceberg.

Even sillier, thats the whole iceberg.

 

[Davy]

I had this with my Gigabyte GA-7VAXP mobo..... turned out to be th
bios battery....! Try a bios re-boot that did the trick for me eve
with the voltage down to 2.3, I got a new battery the next day to b
sure

Not saying this is your problem but certainly wouldn't read the har
drive nor a known good one, funny thing is all the settings in bio
was spot on...even the clock

If you shoudl take this route may be wise to take a note of the bio
settings first, dunno if it was the battery or a dirty spik
(interferance) the mains supply

Dav

 

[Arno Wagner]

Thats radically overstating it.

Have you read the article?

Arno

 

[DRE]

Have you read the article?

Dont need to, I know that YOUR claim is just plain wrong on the
bit you carefully deleted from the quoting and I have restored.

 

[Arno Wagner]

Arno Wagner <[email protected]> wrote:

[...]

Thats radically overstating it.

Have you read the article?

Dont need to, I know that YOUR claim is just plain wrong on the
bit you carefully deleted from the quoting and I have restored.

Why would I "carefully delete"? The old article is there for
everybody to look at. I think you don't understand how usenet
works.

You really should have a look at the article. The 3 out of 30 (o.k., I
re-checked , its actually 3 usable ones out of 27) is entirely
accurate and well explained. Most pleople using commercial Windows
backup programs do not have a reliable backup.

For reference: The ones working reasonable well are "Backit Up",
"Trueimage" and "Genie Backup". Mind that these are not perfect,
the are just adequate. But they are not likely to stab you in the
back, unlike the other programs.

Arno

 

[DRE]

Why would I "carefully delete"?

Ask yourself, you did the deleting. Presumably so the
comments I made after that are no longer in the reply.

The old article is there for everybody to look at.

And its even easier to see those comments
if they arent deleted from the quoting.

I think you don't understand how usenet works.

Been using it for longer than you have thanks.

You really should have a look at the article.

No thanks, I know that YOUR claim about how
many of them are usable is just plain wrong.

The 3 out of 30 (o.k., I re-checked , its actually 3 usable
ones out of 27) is entirely accurate and well explained.

Pity its just YOU running the line about what is
necessary for a particular app to be USABLE.

And you are just plain wrong on that.

Most pleople using commercial Windows backup
programs do not have a reliable backup.

You dont even know what most actually use.

For reference: The ones working reasonable well
are "Backit Up", "Trueimage" and "Genie Backup".

Its completely stupid and pig ignorant to claim
that what comes standard with XP isnt usable.

Sure, some others have definite advantages, but its just plain
silly to claim that what comes standard with XP isnt usable.

I dont need to read that article to know that.

And plenty of others are usable in specific
situations too, like ghost32 for example.

Mind that these are not perfect,

Nothing ever is.

the are just adequate. But they are not likely to
stab you in the back, unlike the other programs.

Mindlessly silly. There are others that
wont do that if you use them properly.

Anyone with a clue checks the log file for example.