current status on HDDs?

[jpdoe]

Re: current status on HDDs?

PS 2: What pcbonez meant when he said that SMART is smart is that the "worst" value is not set in stone. I used to think it coudn't go up, not even if the HDD performance changed for the better. For example: let's say your IDE data cable is broken, and that the SMART CRC errors parameter is is showing a malfunction. In that case, the "worst" value will be low... lower than the threshold. You realize the data cable is the culprit, and you change it. According to pcbonez, given enough time, SMART will realize things changed for the better, and will change the "worst" value to a value above the threshold.
That's what I understood from pcbonez's post, at least.

I didn't observe this phenomenon. What I did observe is that a HDD may give a SMART error now, but if the situation changes, it may not give it in the future.

[jpdoe]

Re: current status on HDDs?

Sorry for the multiple posts.

I agree with dan8139 and kikkoman in that the best test for hdd platter health is a full secuential read monitoring either throughput or access time.
What I do - once in a blue moon - is "dd_rescue /dev/hda /dev/null" or something like it... It's a command present in a bootfloppy called smartlinux... it's a modified dd command that doesn't give up when faced with bad media. The problem is that you have to be in front of the console to monitor the data throughput... and staring at the monitor for 30min is not fun. So next time I'll use one of the utilities mentioned in this thread instead.

[i4004]

Re: current status on HDDs?

kikkoman mentioned particula program...my friend mentioned some program long time ago so i asked him what was it..he said it was this
http://www.ariolic.com/activesmart/index.html
perhaps it's the same program kikko. mentioned.

jpdoe thanks for explanations.
yeap, they surely have a pretty convoluted way of doing things...with all that normalization and equalization...

[jpdoe]

Re: current status on HDDs?

You're welcome, i4004. Thanks for the tip on the software.
I guess the guys that designed SMART took a strange approach to it.

[i4004]

Re: current status on HDDs?

seems they were pouring 1l in 0.5l bottle..
that usually looks strange...

my friend says it was scary how precise this program was in predicting death of his wd 40gb drive...

[zandrax]

Re: current status on HDDs?

@ i4004:
Look in your screenshots for the "Hardware ECC Recovered" parameter: there are millions of recoveries of the so called "unstable" or "pending" sectors (the number is the count of all recoveries, not the number of these sectors).

Unstable are sectors approaching failure but which data can still be decoded with multiple reads and the ECC correction: this, of course, cuts noticeably the transfert rate (disc scanners such as HD Tach show this behaviour as spikes) and isn't a good sign either since those sectors will become unreadeable, "bad", in a near future; moreover, every time the head tries to read these sectors may damage them or the head even more.
Since aging and weak/unstable magnetic coatings cause this and no brand is immune, the only move is backupping all your data to another drive then running a diagnostic which marks these unstable sectors as bad and force the drive to remap them. AFAIK no manifacturer's diagnostic is able to mark, they remap only known bad sectors: however MHDD can mark unstable sectors one by one. Hand marking is slow for more than a hundred sectors: luckily someone developed MakeBad, a script to automatize unstable sectors discovery and marking them.

Suggestion: let the drive reach its working temperature by running for about half an hour, then run MakeBad.
While running MakeBad on a very damaged 2.5" drive (it had over a million unstable sectors), I noticed FreeDOS just crashed with a memory error: I managed to overcome this by reducing the number of sectors MHDD checks during a single run (Maximum LBA sectors).

Good luck.

Zandrax

[i4004]

Re: current status on HDDs?

well, let me at least get some pending sectors first...hehe...

i have copy of 120gb drive that was made recently(on one 160gb drive) and i also have the system(40gb) drive on that 160gb drive but that's a bit older.

new backup strategy will probably be to clone system(40gb) to that 160gb drive, copy contents of 120gb to one external usb disk, and then just put the 160gb disk in and make 120gb partition on it to dump stuff from 120gb drive..though i doubt they'll both die at same time...

i don't really think these drives will start having issues in next year or two...
i'll keep my eye on pending and reallocated, though....

[i4004]

Re: current status on HDDs?

here's the smart of 2 dying hdds(attached images)

both wd and seagate made in 2000...

after certain time(few minutes) wd starts to tick...
then you can't access it etc.
smart shows no pending or realloc. sectors...


seagate has 44 reallocated sectors...if i remember correctly it's problem is that it sometimes works, other times not...

[dood]

Re: current status on HDDs?

Are you only running 640x480?

[i4004]

Re: current status on HDDs?

is that 640x480?

[Per Hansson]

Re: current status on HDDs?

Try some other programs on it and see if they report the same
Seems like the vendors don't use any standardised design for this stuff because I get different readings with different programs
I like HDD Health v2.1 but it's still hit and miss

[i4004]

Re: current status on HDDs?

this was via usb hdd tray and only program supporting it is everest, i think.
will see if hdd health does it, but i doubt....

no, i won't be putting these dying hdds in any of my machines just to see if smart is the same on every interface...

[PCBONEZ]

Re: current status on HDDs?

PS 2: What pcbonez meant when he said that SMART is smart is that the "worst" value is not set in stone. I used to think it coudn't go up, not even if the HDD performance changed for the better. For example: let's say your IDE data cable is broken, and that the SMART CRC errors parameter is is showing a malfunction. In that case, the "worst" value will be low... lower than the threshold. You realize the data cable is the culprit, and you change it. According to pcbonez, given enough time, SMART will realize things changed for the better, and will change the "worst" value to a value above the threshold.
That's what I understood from pcbonez's post, at least.

I didn't observe this phenomenon. What I did observe is that a HDD may give a SMART error now, but if the situation changes, it may not give it in the future.

You have it right.
I can't remember which one, but I talked to either WD or Seagate to learn how that works after I observed a SMART value 'reset' itself from over the limit to under on one of their drives.

.

[Junk Parts]

Re: current status on HDDs?

Well I'm not even in the same book let alone the same page with you savvy people when it comes to this stuff. All I have been tring to do is identify a Dell 4550 mobo. The last tool I tried was Everest Ultimate. Yes this progie give up all sorts of numbers that I don't understand. One place in the stats it gives <DMI> as the mobo ID. It gives the mobo name as Dimension 4550. It lists the mobo chipset as Intel Brookdale i845PE. In the stats for my BIOS string its blank. I'd love to use just the bare 4550 case, and fill it with good stuff. I'm thinking along the lines of a new dual or quad core Intel CPU, Mobo,PSU, RAM, DVD drivers, SATA? Harddrive, and video card. I know I can come up with good cooling solutions if I can use the bare case. After all the Dell Crap is pulled I know there will be room for at least two large case fans. Noise isn't an issue for me as I'm near deaf anyway...LOL

[PCBONEZ]

Re: current status on HDDs?

Noise isn't an issue for me as I'm near deaf anyway...LOL

What?

[Junk Parts]

Re: current status on HDDs?

What? Did somebody say something???