A Colossal HDD Failure

[eccerr0r]

Re: A Colossal HDD Failure

interesting...thought for the most part hard drives shouldn't kill themselves by doing that...

[Dan81]

Re: A Colossal HDD Failure

Do tell how you created exactly one bad sector, this will be interesting

Or is there the possibility of more damage than just that one sector?

I shut off my PC from the PSU switch every time I am done working on it. That's how.

[eccerr0r]

Re: A Colossal HDD Failure

Likely toying with hdparm and it's

--make-bad-sector

option.

but he could just --repair-sector and it's no longer bad because it wasn't really bad to begin with?

[RJARRRPCGP]

Re: A Colossal HDD Failure

I had what I think was a lemon 7200.7 in 2005. I bought it at the Claremont, New Hampshire Wal-Mart, on May 14, 2005, IIRC. I returned it on June 2, 2005, after I was hearing a random squeaking when seeking and utilities expecting HDD failure in less than 1 year, IIRC!

June 2, 2005, was when I came across the Maxtor 6Y060P0, which are virtually unkillable! Those are not the "slimeline" Maxtors.

[BigTroll]

Re: A Colossal HDD Failure

the last seagates I thought were any good and I would see over 50k hours with no problems were the 7200.7 series, I also loved how loud they were.

[momaka]

Re: A Colossal HDD Failure

I junked every Seagate in the house a long time ago, except my dad's old external and a 2004-era 40GB drive, which resides in my Network Bridge Machine.

Eh, don't get me wrong - I still don't mind Seagate HDDs. But all of their newer HDDs (as in, made in the past 10 years), I don't trust. Many seem to develop bad sectors too quickly or just can't read/write anymore, despite SMART passing all parameters with flying colors. And then there's the above 7200.10 HDDs that just plain-straight fail dismally.

Seems the more capacity they've been packing, the less reliable they have become. (Not that this surprises me in any way - more data per given area means smaller and smaller target zones for the heads, and smaller heads riding closer to the surface than ever before.)

Their 7200.9 series is more or less where I draw the line on their reliability, with some of the 7200.9 drives being... well, not so reliable, but still miles better than anything that came after. All in all, though, the 7200.9 line is OK.

7200.7? I'll use and trust any day (not to say I won't keep backups or never expect it to fail, of course.)

Barracuda ATA V, IV, and III? - Same as 7200.7 - old, but reliable workhorses. Too bad they only came in IDE and SCSI variants. At least 7200.7 came in SATA as well.

I also found an old 40 MB (MegaByte) Seagate ST-251 5.25" in the scrap pile at work. Someone removed its top a long time ago just to look at the inside and never put it back on. The thing was dusty as heck when I found it. I gave it a gentle wipe with alcohol. After this, I powered it on, just for shits and giggles and guess what? IT WORKS!!! Seeks, positions, and is ready to work another day... even though I have no system to connect it to. It's a beautiful piece of hardware and makes an awesome boot-up sound. Took it home and definitely keeping it for my retro collection of old PC stuff. No way I could let them get that scrapped.

[TechGeek]

Re: A Colossal HDD Failure

Do tell how you created exactly one bad sector, this will be interesting

Or is there the possibility of more damage than just that one sector?

Likely toying with hdparm and it's

--make-bad-sector

option.

[eccerr0r]

Re: A Colossal HDD Failure

normal Seagate one has 1 bad sector and that was solely caused by me

Do tell how you created exactly one bad sector, this will be interesting

Or is there the possibility of more damage than just that one sector?

[Dan81]

Re: A Colossal HDD Failure

I still have two Seagate drives in my main PC (2x Barracuda ES.2, ST31000340NS - one is Seagate normal, and the other is a IBM branded drive) and both are near brand new condition (I think the normal Seagate one has 1 bad sector and that was solely caused by me). 10 seconds for both to spin up, 5 second each.

As for platters, some WDs were way worse. I've had a 1TB WD Blue (Apple branded no less) LITERALLY pass the seek test with the platters looking the same as momaka's 1.5TB drives.

[TechGeek]

Re: A Colossal HDD Failure

I junked every Seagate in the house a long time ago, except my dad's old external and a 2004-era 40GB drive, which resides in my Network Bridge Machine.

[momaka]

Re: A Colossal HDD Failure

I'm back with another colossal HDD failure.
Same brand and model (Seagate 7200.11, model ST31500341AS.)


This hard drive came from the same bin at work where I got the previously (badly) failed one. However, when I powered this one on my test PC, it actually sounded as if it's trying to seek / calibrate, and I could swear I almost heard it do it a few times. But after numerous power ups/downs and tries, while the HDD was running, all of a sudden I heard a loud clunk and the HDD jumped violently. After this, I heard the HDD try to spin back up, followed by that awful grinding noise again.

What are the chances this one developed a head crash? Pretty HIGH! So without further thought I opened it up, and sure enough was greeted with the same mess again:



Want the data back? I'm sorry, but you will have to pluck those magnetic bits off one by one from the head arm magnet and the filter:


I think Seagate should write on the label of these HDDs "Certified Trash".

I got several more 7200.10 and 7200.11 models from that same bin at work (though smaller capacities), and they all have various problems. Mostly bad sectors and slow reads. Not a single one over 50k hours, IIRC.

[Curious.George]

Re: A Colossal HDD Failure

With 56K and earlier, this was difficult, compared to broadband with an Ethernet interface. The usual dial-up configuration, didn't use an Ethernet interface at the client end.

I had a little box that had a PLIP i/f on one end and serial to a modem on the other. (PLIP gave me a wider pipe to my host so I could also "do stuff" on the little box without impacting bandwidth between it and my host)

The usual, was to use solely a software firewall.

Back in November, 2003, IIRC, I think I forgot to install the firewall software and presto, I got attacked!

Yup. Like AIDS -- it only takes *one* unprotected experience...

[RJARRRPCGP]

Re: A Colossal HDD Failure

Don't put ANY machine directly on a public network. You should always have SOME sort of appliance between you and the Real World. Once you have an appliance in place, you can "bend the rules" and leave all ports CLOSED -- yet still NOTICE when an attempt is made to access one.

Ever make a phone call, let it ring twice, then hang up? Then, call back and know that your call will be answered? Because you've transferred "information" to the called party in that first, aborted call even though they never answered; they now know that another call coming in on the heels of the first will likely be *you*.

Any other caller, on the other hand, doesn't even know if the phone "made an audible ringing sound" -- there might not be a phone connected to that line! So, a caller can let it ring 100 times and get no information from that effort.

With 56K and earlier, this was difficult, compared to broadband with an Ethernet interface. The usual dial-up configuration, didn't use an Ethernet interface at the client end.

The usual, was to use solely a software firewall.

Back in November, 2003, IIRC, I think I forgot to install the firewall software and presto, I got attacked!

[Curious.George]

Re: A Colossal HDD Failure

That was Microsoft's fault, for making an artificial requirement that port 135 be open, IIRC. Also the same for port 445, IIRC. At least for XP, if not also later, possibly 2000 as well. Possibly all NT-based Windows versions since NT 5, if not NT4 as well!

IIRC, no such animal for Windows 95-based OSes.

Don't put ANY machine directly on a public network. You should always have SOME sort of appliance between you and the Real World. Once you have an appliance in place, you can "bend the rules" and leave all ports CLOSED -- yet still NOTICE when an attempt is made to access one.

Ever make a phone call, let it ring twice, then hang up? Then, call back and know that your call will be answered? Because you've transferred "information" to the called party in that first, aborted call even though they never answered; they now know that another call coming in on the heels of the first will likely be *you*.

Any other caller, on the other hand, doesn't even know if the phone "made an audible ringing sound" -- there might not be a phone connected to that line! So, a caller can let it ring 100 times and get no information from that effort.

[RJARRRPCGP]

Re: A Colossal HDD Failure

Having a slow link doesn't *prevent* infection. Rather, it makes it more obvious when someone is stealing your bandwidth for uses that you haven't approved (e.g., siphoning your data, capturing your keystrokes, running a botnet, etc.)

If you already have an "open" connection to your server, then the connection can be "detected". OTOH, if you deliberately DON'T open a connection to your server until you KNOW the "caller" is someone you should let in, hackers just see you as a "black hole".

E.g., if I have a machine at W.X.Y.Z but you can't tell if my machine is plugged in, turned on, or destroyed in a recent fire, then you're not going to hang around waiting to see if it magically appears. Move on to some other IP address that actually *looks* to be present!

That was Microsoft's fault, for making an artificial requirement that port 135 be open, IIRC. Also the same for port 445, IIRC. At least for XP, if not also later, possibly 2000 as well. Possibly all NT-based Windows versions since NT 5, if not NT4 as well!

IIRC, no such animal for Windows 95-based OSes.

[Curious.George]

Re: A Colossal HDD Failure

I remember a virus back in 2003, where you will get infected quickly on 56K!
A firewall was required to prevent the exploit. It was "Blaster". (a.k.a. msblast.exe)

How did I know that I got exploited? The remote procedure call (RPC) service crashes.

Having a slow link doesn't *prevent* infection. Rather, it makes it more obvious when someone is stealing your bandwidth for uses that you haven't approved (e.g., siphoning your data, capturing your keystrokes, running a botnet, etc.)

If you already have an "open" connection to your server, then the connection can be "detected". OTOH, if you deliberately DON'T open a connection to your server until you KNOW the "caller" is someone you should let in, hackers just see you as a "black hole".

E.g., if I have a machine at W.X.Y.Z but you can't tell if my machine is plugged in, turned on, or destroyed in a recent fire, then you're not going to hang around waiting to see if it magically appears. Move on to some other IP address that actually *looks* to be present!

[RJARRRPCGP]

Re: A Colossal HDD Failure

I remember a virus back in 2003, where you will get infected quickly on 56K!
A firewall was required to prevent the exploit. It was "Blaster". (a.k.a. msblast.exe)

How did I know that I got exploited? The remote procedure call (RPC) service crashes.

[Curious.George]

Re: A Colossal HDD Failure

exactly, i subscribed to a slower connection to save money. 200mbps is fast enough for me. this isnt a dick wagging contest!

when torrenting, im more of a hit and run torrenter. i just spend a few days downloading and re-seeding stuff then im gone. makes it hard to track me down and my activities!

Any sort of "large data transfer" can almost always be spread out over time (the exception being HUGE torrents that aren't seeded for long periods of time). So, just choose that "time" to be one when you won't really care to be waiting (e.g., asleep and totally oblivious of how long it might take; your MACHINE has infinite patience!).

Note that you are always at the mercy of the data source's pipe; someone seeding at 100Kbps is going to take just as long to DL a given torrent over 10Gbps and 10Mbps links! :> Ditto pulling material off of an FTP server or web site.

[and who the hell wants to be MS's bitch distributing updates to peers "for free"?]

The downside of fat pipes is that your machine can be targeted, infected and compromised before you even notice the surge in traffic!

I have a (hidden/secret) back door to access my servers. The pipe is deliberately constricted so that it would take *ages* to get anything in or out through that connection. The connection is normally used for administration -- which is usually a very low bandwidth activity (CLI). So, there is no "cost" to this restriction in normal use!

The appliance that stands watch over that entryway is particularly sensitive to attempts to hammer at it, repeatedly -- it KNOWS that I would never behave like that. So, anyone trying to make repeated attempts to gain entry using the wider pipe leading into the house effectively reveals their malevolent intentions BEFORE they have had a chance to even knock on the door!

And the appliance just "dummies-up".

[ChaosLegionnaire]

Re: A Colossal HDD Failure

exactly, i subscribed to a slower connection to save money. 200mbps is fast enough for me. this isnt a dick wagging contest!

when torrenting, im more of a hit and run torrenter. i just spend a few days downloading and re-seeding stuff then im gone. makes it hard to track me down and my activities!

[eccerr0r]

Re: A Colossal HDD Failure

My 2TB disks are getting up there in hours, have many of them around the 60K mark now. Getting close to my personal record of the 120G disks that are around 70K hours and still working. Currently the 2T disks are Hitachi/Toshiba, Seagate, WD Green, and a WD Black (I think). Except for the green, the other disks are in some RAID. The WD black was a refurb and also the disk that has the highest potential for failing first...grr... The green has had its expected head load/unload cycles completely exhausted, luckily it's still working. This power save measure has been disabled.

Though this is NOT necessarily accurate, the main partition of that 2TB Green reports
Lifetime writes: 15 TB
and it's probably the disk I've had most I/O to, mostly because it's a PVR disk. Seems the next most abused disk I have is one of my SSDs that I've done 8TB writes to. I have a feeling that having swap on this disk is part of the reason why it's fairly high though the wear counter has barely ticked yet despite the large number of writes (it should tick soon...)

Oddly enough the number of bytes written differs from the filesystem (8TB) and from SMART (12TB) differ on that SSD. What gives. Probably metadata may not be included? No idea.

Anyway, the disks that I got failures on the surface like the pictured ones was apparently due to orientation or contamination. An old 40M WD disk and some other disk I no longer remember...