Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Kevin Wood]

Does drive make matter?

Yep. Some are better than others. You'd need to ask someone who uses lots of them for current recommendations.

Some drives will last for years, some will pack up after a year or two. Thing is, they're cheap relative to the data you have on them so you may as well change them well before they get into the steep part of the MTBF curve! My experience is that, after 3 years, it's not worth keeping them as they may fail at any time (they may go on for another 10 years, of course), especially if you're rebuilding the system anyway.

I certainly wouldn't rely on a 10 year old drive in a server  

Kevin

[TheBoy]

Does drive make matter?

I have used lots of 4 year old drives with no issues.

We had a server at work with a 10 year old drive, started as the original SYS volume and just stayed there update after update

No, they are all much or a muchness.

After about 3yrs, the likelyhood of failure increases significantly.  I have some drives that are still working despite being 10+ yrs old, but then I've had several fail at the 3 - 5yr age.  If you don't get ELF on them, its likely they are good for 3 or 4 yrs. Much longer is too big a risk for me (for data).

As for a 10yr old drive in a server, thats a bit daft.  Assuming you have no data, and you won't lose money in any restore, the heat/energy used for the space available (2 - 4Gb?), and the speed of the device, is it worth it, when a modern 320G is only £50 - £70, and will use less power, generate less heat, be faster, and should be more reliable.....

[TheBoy]

I certainly wouldn't rely on a 10 year old drive in a server  

Esp in a kNitWare server (assumption based on server age, and SYS volume). NetWare 3/4/5 servers are not reknown for their recoverablility when corrupted - I've been there several times

[megaomega123]

Any unusual noises from the Hard Drive?

If it's starting to go bad I'd be inclined to get a new hard drive ASAP and copy everything over to it while you still can! Modern hard drives have internal strategies to remap bad blocks and this is transparent to the OS. If the OS is starting to see hard drive issues it's bad news. Could still be the drive controller or another hardware issue though.

How old is the drive? If it's more than 2 or 3 years old it's on borrowed time.


Kevin

Yes it is, Kevin. It's a doddle to physically change the drive, but copying everything over is something I do not relish (gone are the days of COPY *.*!). Can anyone recommend a good ghost utility to do this? I've got three drives, but I think it is my main C drive with programs on it that's failing.  

If you have 3 drives installed then you can make a back up file on 1 of the other drives and drag my docs, my pics etc into it. If you use outlook you can back up your setting by copying 2 files and if you use firefox again you can backup your profile easily.
You could always try a simple chkdsk. Go to My computer, right click on the affected drive and select properties. Go to the Tools tab and click check now, tick both tick boxes then start.

[Martin_1962]

I certainly wouldn't rely on a 10 year old drive in a server  

Esp in a kNitWare server (assumption based on server age, and SYS volume). NetWare 3/4/5 servers are not reknown for their recoverablility when corrupted - I've been there several times

We just kept adding drives to it - until SYS went bang and we lost the O/S

So it was rebuilt as a 6.x

[sounds2k]

the other thing - depending on your system - you could buy 2 hard disks instead of 1, then set them up as a RAID-1 array - this copies everything on the one drive to the other in real time - which should protect against hardware failure. It's the route I went down after nearly losing all my data due to a dodgy (IBM deskstar, I mean deathstar) hard drive ...

A lot of systems have the ability to do RAID-1 built into the main board, if not an add-on card should cost about £30.

However, it's no substitute for a backup on an external device (another hard drive in a USB2 or firewire case is good - even better if only connected to backup or restore data), ideally use both - as if you were to get a virus it could delete the data on one drive - which would then get deleted on the other as well ... but if you've got a backup on a hard disk which isn't even connected then you should be safe  

I've also had a fair bit of success with acronis trueimage for backing up and restoring the system, it worked on my new motherboard which has a nvidia chipset whereas ghost 9 refused to load the supposedly certified driver  

[Nickbat]

I use Drive C for my program files, Drive D for my data and Drive E for data backup and other stuff.

I did a CHKDSK on Drive C. Unfortunately, it said it couldn't gain exclusive access to the drive and could only run on rebooting. This was fine, except that having waited an hour for it to check, the final report of what it found flashed on the screen for about 1 nanosecond before it continued to reboot. So, I don't know if it found bad sectors and repaired them or what. :-/

Ho hum. I'll next do a CHKDSK on my Drive D.

I KNOW I need to buy a new main HDD and reload all my software. It's just one of those jobs (like the new cambelt and service on the car) that I really must get round to doing..... Oh, and my VAT return, too.

[TheBoy]

Martin Imber - inevitable. Not the hassle of doing it, but lost revenue whilst the system is down. Thats why you don't run old drives!

Sounds - the motherboard RAIDS are pretty naff, mostly software driven, with the hassles and performance overheads that involves, and the drivers tend to be flakey.  Decent RAID hardware can be bought cheaply, as you say though (bit more than £30 for the good stuff).  Also the drivers tend to be Windows, and Linux if you're really lucky. Novell, Solaris etc don't get a look in.  Software RAID by the OS is an option, but tends to be poor performance under most OSes, and troublesome to boot of the other half of the mirror.

[Martin_1962]

Martin Imber - inevitable. Not the hassle of doing it, but lost revenue whilst the system is down. Thats why you don't run old drives!

Sounds - the motherboard RAIDS are pretty naff, mostly software driven, with the hassles and performance overheads that involves, and the drivers tend to be flakey.  Decent RAID hardware can be bought cheaply, as you say though (bit more than £30 for the good stuff).  Also the drivers tend to be Windows, and Linux if you're really lucky. Novell, Solaris etc don't get a look in.  Software RAID by the OS is an option, but tends to be poor performance under most OSes, and troublesome to boot of the other half of the mirror.

Good job we had two servers then with everything on both

[TheBoy]

Martin Imber - inevitable. Not the hassle of doing it, but lost revenue whilst the system is down. Thats why you don't run old drives!

Sounds - the motherboard RAIDS are pretty naff, mostly software driven, with the hassles and performance overheads that involves, and the drivers tend to be flakey.  Decent RAID hardware can be bought cheaply, as you say though (bit more than £30 for the good stuff).  Also the drivers tend to be Windows, and Linux if you're really lucky. Novell, Solaris etc don't get a look in.  Software RAID by the OS is an option, but tends to be poor performance under most OSes, and troublesome to boot of the other half of the mirror.

Good job we had two servers then with everything on both

Clustered?

[Martin_1962]

Martin Imber - inevitable. Not the hassle of doing it, but lost revenue whilst the system is down. Thats why you don't run old drives!

Sounds - the motherboard RAIDS are pretty naff, mostly software driven, with the hassles and performance overheads that involves, and the drivers tend to be flakey.  Decent RAID hardware can be bought cheaply, as you say though (bit more than £30 for the good stuff).  Also the drivers tend to be Windows, and Linux if you're really lucky. Novell, Solaris etc don't get a look in.  Software RAID by the OS is an option, but tends to be poor performance under most OSes, and troublesome to boot of the other half of the mirror.

Good job we had two servers then with everything on both

Clustered?

No migrated, we were building a Windows server for testing and it was already a backup server - pity it is slower!

The 6.x is used for testing now

[Markjay]

There I was reading markjay's excellent description of 2.6 plenum removal, when ZAP!. The blue screen of death appeared with about three lines of code at the top and the words: Preparing to dump physical memory.

I powered off and rebooted but what the fubar was that all about?


Nick
Using Win 2k Pro

Nothing to do with me mate  


 

[Nickbat]

OK, so after a week without problems, the BSOD came back today. (Again, I was browsing the forum!)

This time, I noted the down the message in full:

***STOP: 0x00000000, 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0x00000000)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory


Does that provide any pointers?

[TheBoy]

OK, so after a week without problems, the BSOD came back today. (Again, I was browsing the forum!)

This time, I noted the down the message in full:

***STOP: 0x00000000, 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0x00000000)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory


Does that provide any pointers?

Duff driver normally. That may not necessarily be a hardware driver, but also 'virtual drivers' as often used by software firewalls and AV...

Did it give the process name that it happened in?

See also http://kadaitcha.cx/stop_err.html#0x000000D1

[Nickbat]

OK, so after a week without problems, the BSOD came back today. (Again, I was browsing the forum!)

This time, I noted the down the message in full:

***STOP: 0x00000000, 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0x00000000)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory


Does that provide any pointers?

Duff driver normally. That may not necessarily be a hardware driver, but also 'virtual drivers' as often used by software firewalls and AV...

Did it give the process name that it happened in?

See also http://kadaitcha.cx/stop_err.html#0x000000D1

No, it didn't. Strange though, that on the only two occasions this has happened I have been browsing OOF. Last time it was the plenum removal in the maintenance guides, this time it was Mark's autobox guide!!

I know it has nothing to do with OOF, but how's that for a coincidence?