- Bobb - said:
Thanks guys.
When I checked, I do see that internal are drives 0+1 so the "issue" was
with my fairly new external WD 1 terabyte drive ??
Before posting, I checked that error at the Microsoft link and there is said
" if not the system disk then ignore the error". How useless an info message
is that ? IS there a problem or not ? Should I return under warranty ??
In general I think all of MSFT's troubleshooting is related to " is it our
fault?" If not, then .... we don't care.
Usually that is my backup drive and is turned off .
For now I'll live with it and see if it happens again .If so, then I'll try
to get the data copied to yet another drive. ( when I was shopping I WAS
going to get a Seagate, but reviews said WDC was better. agrrrrrrhhh.)
I would recommend going to the Western Digital site and downloading
DataLifeguard tools for disk testing. I don't know what restrictions
exist on that software, like what interfaces it doesn't work on.
Perhaps if you run a diagnostic, it will confirm there is some
kind of problem. The download page, specifies tools as a function of
the drive model, so presumably they'll list a test tool that is
relevant.
The OS is written by Microsoft, such that kernel operations remain
responsive. Now, if you have some software present, that violates
any of the real time properties, or inserts shims or the like in
a software stack, that can change the properties in a way not
anticipated by the OS designers. An example of intrusive software
might be some AV software. Googling on "Delayed Write" failures,
from the past, might dig up software situations, where some
third party addition has upset proper OS operation.
Some hardware operations on disks, have longer timeout constants than
perhaps the driver is willing to wait. Say the disk will attempt for
20 seconds, to read a flaky sector. And the OS disk driver will only
wait 5 seconds for the disk to respond. That could result in log
entries being generated, and perhaps even a consistency problem.
There are some disks, where a parameter is changed, to make that
less of an issue in certain situations. Western Digital makes RE
drives, where for twice the price of a regular disk, they change
the timeout constant to somewhere around 5 to 7 seconds. The purpose
of that, is so a severe disk issue, won't cause a RAID array to go
offline. (The disks are intended for usage in RAID arrays.) At one
point, some of the WD disks, supported the usage of a utility, to
flip the timeout constant, without paying an exorbitant premium
to WD, but they closed that loophole.
So there could be physical reasons for this, such as a bad brand
new disk. Or, it could be software induced (for more reasons than
I know about).
I'd start with a backup first, leaving the drive powered off
until you're ready to run the transfer. Connect a replacement disk,
and make sure all your data is safe.
After that, try a chkdsk, and see if any problems are detected
with the file system(s) on that disk. If there is a persistent
problem, and it is inserting problems in the file system structures,
the longer you leave it, the harder it would be for chkdsk to
repair it.
Shut down, and run the WD diagnostics. (Some diagnostics use read
only testing, with options to do more destructive write testing.
I usually just stick with read testing.)
The SMART statistics kept by the drive, have a few key parameters
that tell you about drive health. I couldn't decode that info,
if my life depended on it
But one key parameter, is the
"Pending" count. That tells you, how many suspect sectors the
drive has detected. On the next write attempt to those sectors,
the drive will have to decide whether that sector is dependable
or not. If the next write fails to work, then the sector is
spared out. Seeing a persistent pending count, might be a
sign of trouble to come. I don't think I've seen any "Pending"
accumulate on drives here, when I've looked at the SMART stats.
The file system, also has mechanisms for tracking "bad clusters",
and that is a search term you could try, for more information
on where that kind of thing might be logged. For me, the
logging of physical disk problems is more important, as I
spook easily, and at the first sign of trouble, I install
a new disk. It's one of the reasons I have so many
spare disks sitting on a table in the junk room
Paul
