Pendiong Sector Count Increasing (Rapidly?)

[(PeteCresswell)]

Recently had a drive where the bad sector count was going out of
sight and I swapped it out for another drive.
("Hard Disk Sentinel: Drive Is Failing... Now It's OK?")

Now the replacement drive's "Pending Sector Count" seems tb
increasing rather rapidly as in http://tinyurl.com/bghyavd

This drive is formatted with 64k sectors and is used to record
television shows.

There's another drive on the same mobo that does the same thing
and there have been no problems with it.

What's got me going is the original drive's bad sectors going out
of sight, and then the replacement drive's Pending Sector Count
increasing so rapidly.

Unencumbered by any knowledge, I can think of four explanations:

- I am mis-interpreting the log and the Pending Sector count
is within reason.

- Since the drive is being recorded to so heavily compared to
how it was being used in it's prior life, it is natural to
uncover a lot of weak sectors at first

- I've managed to put two marginal drives in a row on that
controller.

- There's something goofey with the controller and it's hosing
the drive bit-by-bit.


Do any of these get anybody's vote?

Other explanations?

 

[Paul]

Recently had a drive where the bad sector count was going out of
sight and I swapped it out for another drive.
("Hard Disk Sentinel: Drive Is Failing... Now It's OK?")

Now the replacement drive's "Pending Sector Count" seems tb
increasing rather rapidly as in http://tinyurl.com/bghyavd

This drive is formatted with 64k sectors and is used to record
television shows.

There's another drive on the same mobo that does the same thing
and there have been no problems with it.

What's got me going is the original drive's bad sectors going out
of sight, and then the replacement drive's Pending Sector Count
increasing so rapidly.

Unencumbered by any knowledge, I can think of four explanations:

- I am mis-interpreting the log and the Pending Sector count
is within reason.

- Since the drive is being recorded to so heavily compared to
how it was being used in it's prior life, it is natural to
uncover a lot of weak sectors at first

- I've managed to put two marginal drives in a row on that
controller.

- There's something goofey with the controller and it's hosing
the drive bit-by-bit.


Do any of these get anybody's vote?

Other explanations?

Use HDTune health screen, and see if the data is interpreted the same way.
Same counts and so on. Just in case the definition of some field is
wrong (like, power on hours is being used as the Pending Sector
field or something).

http://www.hdtune.com/files/hdtune_255.exe

Paul

 

[Bob F]

Recently had a drive where the bad sector count was going out of
sight and I swapped it out for another drive.
("Hard Disk Sentinel: Drive Is Failing... Now It's OK?")

Now the replacement drive's "Pending Sector Count" seems tb
increasing rather rapidly as in http://tinyurl.com/bghyavd

This drive is formatted with 64k sectors and is used to record
television shows.

There's another drive on the same mobo that does the same thing
and there have been no problems with it.

What's got me going is the original drive's bad sectors going out
of sight, and then the replacement drive's Pending Sector Count
increasing so rapidly.

Unencumbered by any knowledge, I can think of four explanations:

- I am mis-interpreting the log and the Pending Sector count
is within reason.

- Since the drive is being recorded to so heavily compared to
how it was being used in it's prior life, it is natural to
uncover a lot of weak sectors at first

- I've managed to put two marginal drives in a row on that
controller.

- There's something goofey with the controller and it's hosing
the drive bit-by-bit.


Do any of these get anybody's vote?

Other explanations?

Do a full RAM test. Any errors? (Memtest+)

 

[Yousuf Khan]

Recently had a drive where the bad sector count was going out of
sight and I swapped it out for another drive.
("Hard Disk Sentinel: Drive Is Failing... Now It's OK?")

Now the replacement drive's "Pending Sector Count" seems tb
increasing rather rapidly as in http://tinyurl.com/bghyavd

A pending sector is not a completely bad sector. It is a sector that is
still readable, but it may be having some trouble being read quickly
enough (i.e. need to reread it multiple times). Usually this is fixed by
rewriting to that sector, thus the final sector status is "pending" a
write to it. It's only after writing to it that the sector is either
sent to the bad sector pool, or back to the good sector pool.

If you'll notice from your log, your pending sector count is not
actually skyrocketing, it's actually whipsawing; it's going up and down.
At one point, the pending sectors is upto 39-40, and then the next entry
it's gone back down to 15-16. This is the drive removing the sectors
from the pending sector count automatically, either to the good pool or
the bad pool. If it had moved them to the bad pool then you'd see a
message in the log about it moving sectors to the bad pool, but since
you don't see those, that means they were moved to the good pool.

If the pending sector count bothers you then HD Sentinel has a
destructive write tester feature that'll write to every sector on the
drive and thus remove all of the pending sectors. But of course you'll
lose all of your data on it, unless it's backed up, because this is like
a secure hard drive wipe.

This drive is formatted with 64k sectors and is used to record
television shows.

There's another drive on the same mobo that does the same thing
and there have been no problems with it.

What's got me going is the original drive's bad sectors going out
of sight, and then the replacement drive's Pending Sector Count
increasing so rapidly.

It sounds like maybe you have an underpowered PSU. This often happens
when you start adding a lot of hard drives to a system, its power
requirements rapidly escalate beyond what you originally thought was
sufficient. My system has grown to 7 internal storage drives (1 optical,
1 SSD, and 5 HDD's), and my PSU has also had to grow from a 400W to a
750W now. Without enough power going to the hard drives a lot of reads
and writes may fail and it could get interpreted as bad sectors, pending
sectors, etc. And you'll find that some hard drives are just fine and
have no problems, but others are acting flaky. That's because the drives
that may be fine, may have a better connection to the PSU, than the
others and may in fact be robbing some of the power away from the
others. Try this PSU calculator to see if you have a sufficient PSU for
the job:

eXtreme Power Supply Calculator
http://extreme.outervision.com/psucalculatorlite.jsp

So it's entirely possible that that the original drive is just fine. The
drive may periodically re-evaluate its own bad sectors and find that
they're fine, and bring them back into the good pool. Not all drives do
this, but I've found a few that did this, all of a sudden they have no
bad sectors on them anymore.

Yousuf Khan

 

[(PeteCresswell)]

Per Yousuf Khan:

A pending sector is not a completely bad sector...
...your pending sector count is not actually skyrocketing,
it's actually whipsawing; it's going up and down.

Great stuff!.... thanks. Now I know more than when I started
this little journey.

It sounds like maybe you have an underpowered PSU. ...
Try this PSU calculator to see if you have a sufficient PSU for
the job:

eXtreme Power Supply Calculator
http://extreme.outervision.com/psucalculatorlite.jsp

That was interesting. It came up with a recommended 480 watts
and my PSU is only 500.

Per your observations, though, it seems possible that the current
messages from Hard Disc Sentinel are not indicative of any
problem at all.

OTOH, if problems seem tb developing, I think I'll move to a
bigger PSU - like a 600 or 750 - just on GPs. Never hurts to
have a spare PSU laying around...

So it's entirely possible that that the original drive is just fine.

Ran Seagate's test utility on it and the utility threw a couple
of different error codes that (hopefully) qualified the drive for
warranty replacement. It arrived at their facility yesterday...
In it's previous life, it was part of the RAID array on my NAS
box... so maybe it got more than it's share of activity there.

Not all drives do
this, but I've found a few that did this, all of a sudden they have no
bad sectors on them anymore.

I ran HD Tune's "Long" test on it came up with 15 "Damaged
Blocks" - whatever they are in HD-Tune-speak. Even though it's
sounding like a red herring, I'll give it another run in a month
or so just to see how it develops.


Thanks again for the consciousness-raising.

 

[glee]

Per Yousuf Khan:


Great stuff!.... thanks. Now I know more than when I started
this little journey.



That was interesting. It came up with a recommended 480 watts
and my PSU is only 500.

Per your observations, though, it seems possible that the current
messages from Hard Disc Sentinel are not indicative of any
problem at all.

OTOH, if problems seem tb developing, I think I'll move to a
bigger PSU - like a 600 or 750 - just on GPs. Never hurts to
have a spare PSU laying around...




Ran Seagate's test utility on it and the utility threw a couple
of different error codes that (hopefully) qualified the drive for
warranty replacement. It arrived at their facility yesterday...
In it's previous life, it was part of the RAID array on my NAS
box... so maybe it got more than it's share of activity there.




I ran HD Tune's "Long" test on it came up with 15 "Damaged
Blocks" - whatever they are in HD-Tune-speak. Even though it's
sounding like a red herring, I'll give it another run in a month
or so just to see how it develops.


Thanks again for the consciousness-raising.

Current Pending Sector Count S.M.A.R.T. attribute
http://www.ariolic.com/activesmart/smart-attributes/current-pending-sector-count.html

 

[Yousuf Khan]

Per Yousuf Khan:


Great stuff!.... thanks. Now I know more than when I started
this little journey.



That was interesting. It came up with a recommended 480 watts
and my PSU is only 500.

Prior to this 750W PSU that I have, I had an intermediate 600W PSU,
which seemed to be adequate to do the job, at least on paper. However,
that PSU was built to an older standard, where the the +5V output were
split into multiple rails (4 of them). At one time, there was an EU
safety regulation about PSU rails never exceeding 20A (the standard may
still exist). So if you wanted to have more than 100W (5V * 20A)
available through the 5V channel, you needed to split it into multiple
rails. So usually one or two of these rails would be reserved for the
CPU and motherboard, and a third for a GPU, and one lone rail for
everything else in the system, including hard drives. So the system
seemed outwardly stable, as the CPU and GPU were more than adequately
fed, but the hard drives were starving, having to share their power with
other hard drives, and every other peripheral in the system, including
cooling fans, LED's, etc. So I'd end up with situations where a hard
drive would just end up disconnecting by itself randomly, as if it it
didn't exist inside the system; the reboot would bring it back of course.

That was only an EU regulation, and never a requirement in North
America, but they probably wanted to keep the same models available in
the EU and NA to save on costs. But I think PSU mfg's now simply make a
separate line of PSU's for North America, which now have a monster
single-rail. This makes power distribution much easier and much more
no-brainer. So the new 750W is of the monster-rail variety (by itself,
it probably has 100A flowing through it), and everything gets fed from
the same pool of power.

Per your observations, though, it seems possible that the current
messages from Hard Disc Sentinel are not indicative of any
problem at all.

I wouldn't say HDS is indicating nothing, if the power situation is left
unresolved long enough, then there may indeed some real permanent
physical damage.

Yousuf Khan