WD Raid Edition Drives in Non-Raid Desktop

[John G.]

How big of a no-no is it to use Western Digital's RE Caviar drives in a non-RAID setting? I know the TLER feature of these drives make them "unsuitable for desktop use," but how much wiggle room is there on that?

I ask because I have a file server that I'm planning to convert over to RAID 5 in the near future, but am not ready to make the (large) initial investment in a good RAID controller and replacing all my IDE Caviar SE drives with SATA RE ones. But I'm definitely going to need more storage capacity before the switch to RAID, and I'd rather not keep buying IDE Caviar SE drives that I know I'm just going to have to resell in a few month's time. So can I go ahead and buy a SATA RE drive (or two) and just stick them in the system without RAID until I have the capital to build the file server I really want?

How often do hard drive errors occur? How probable is it that they would take longer than seven seconds to correct? And what would happen in a desktop setting if the correction did not complete in time and there was no RAID card to pick up the slack?

[draksia]

Hard drives for servers are designed with the assumption that there is a RAID controller present and some coordination of error management must occur. WD has delivered that coordinated error management in the form of Time Limited Error Recovery (TLER). TLER-capable hard drives will perform the normal error recovery, and after 7 seconds, issue an error message to the RAID controller and defer the error recovery task until a later time. With coordinated error handling, the hard drive is not dropped from the RAID array, thereby avoiding the entire RAID recovery, replacement, rebuild, and return experience. The error handling is further coordinated between the TLER-capable hard drive and the RAID card. The TLER-capable drive will respond without waiting on the error to be resolved. RAID cards are very capable of handling this with a combination of parity protection and journaling. The RAID card flags the error in the error log and proceeds to deliver data using parity protection until the drive retries its own error recovery and corrects the error. This is quite similar to error management proven in SCSI-RAID for many years. It is important to realize TLER-capable hard drives should not be used in non-RAID environments.


That is specifically what Western Digital says. Unfortunately WD only list the unrecoverable error rate so it is hard to determine how often you would get an error. Because of the TLER feature works though a message will be present in your system log that says if the drive encountered an error. Because of the relatively low load placed on desktop drives with the TLER it is still likely that the drive repair the error no problem. My advice is that while it is not great you can probably get away with using them especially for a limitted time.

[texuspete00]

Tech report's end of year piece gave the RE2's drive of the year and they said go ahead and use it single in there I believe.

[John G.]

Okay, so now there's a problem.

I purchased the RE drive from ZZF, and since I didn't have a motherboard with SATA ports, a cheap controller card (http://www.newegg.com/Product/Product.asp?Item=N82E16815124022) from Newegg. I installed it last night, the computer had no trouble detecting it, initializing the hard drive, formatting the drive, and putting about 250GB of data onto it.

Now, funny thing is, whenever I try to access the new drive, I get serious lag. Like, the computer freezes up for about two-three seconds every five seconds or so. Mouse stops moving, programs don't respond (in that two-three second window). Very very annoying. I also can't stream video off the new drive because of this problem. Transfer rates to the new drive are abysmal because of it. So it seems to affect all read/write operations of the drive.

I can't recall if it's been doing this from the start, but I suspect it has.

Any ideas what might be causing this? I'm concerned that the power supply might not be having such a good time since I installed the new drive. The RE makes for drive number five in an old PIII Dell Optiplex system (not exactly enterprise class PSU). But I've never heard of a taxed power supply resulting in symptoms like this.

[hokatichenci]

Is DMA on? Are you having any IRQ conflicts?

[John G.]

No IRQ conflicts, and I can't figure out how to check for DMA. The drive properties doesn't have a "Settings" tab on it-- either because it's a SATA drive or because it's on a controller card.

[hokatichenci]

Assuming SATA can do DMA (I dunno honestly), it might be a controller card issue. I have a Promise IDE card that can't do DMA for some reason and I never was able to figure out why. Performance was total crap when moving files off it, and it had problems streaming high quality videos sometimes because the bandwidth was just choking. If you have a lot of other cards in the system you might also just be choking the bandwidth of the bus. Could you post full specs on the machine?

[hity645]

So its a no-no to use RE drives in non raid setups?

[John G.]

Assuming SATA can do DMA (I dunno honestly), it might be a controller card issue. I have a Promise IDE card that can't do DMA for some reason and I never was able to figure out why. Performance was total crap when moving files off it, and it had problems streaming high quality videos sometimes because the bandwidth was just choking. If you have a lot of other cards in the system you might also just be choking the bandwidth of the bus. Could you post full specs on the machine?

Someone had suggested the same thing to me, that I might be overcrowding the PCI bus. The machine is as follows:

Dell OptiPlex GX110, Made in 2001
1GHz PIII w/ 512MB PC100 SDRAM
NVidia MX4000 128MB Video Card (on PCI bus)
Standard Sound Card (on PCI bus)
SATA/IDE Combo Controller (on PCI bus)

The machine has five hard drives in it, an 80GB and 320GB on the mobo's primary IDE channel, two 320GB on the secondary IDE channel, and the 400GB on the controller card.

I'll do some tests later tonight to find out if it is the controller card/video card combo saturating the PCI bus. Any thoughts in the mean time?

[hokatichenci]

Someone had suggested the same thing to me, that I might be overcrowding the PCI bus. The machine is as follows:

Dell OptiPlex GX110, Made in 2001
1GHz PIII w/ 512MB PC100 SDRAM
NVidia MX4000 128MB Video Card (on PCI bus)
Standard Sound Card (on PCI bus)
SATA/IDE Combo Controller (on PCI bus)

The machine has five hard drives in it, an 80GB and 320GB on the mobo's primary IDE channel, two 320GB on the secondary IDE channel, and the 400GB on the controller card.

I'll do some tests later tonight to find out if it is the controller card/video card combo saturating the PCI bus. Any thoughts in the mean time?

If you're only using video/sound on the PCI bus (ie, not using a gigabit nic/many other controllers) I highly doubt you could saturate it. I was able to saturate the pci bus on my old system when I ran a NIC, audio, and 2 controller cards. It wasn't an easy task but I'm thankful for Pci-e ;) I'm putting my bet on something DMA related with the controller card.

[John G.]

If you're only using video/sound on the PCI bus (ie, not using a gigabit nic/many other controllers) I highly doubt you could saturate it. I was able to saturate the pci bus on my old system when I ran a NIC, audio, and 2 controller cards. It wasn't an easy task but I'm thankful for Pci-e ;) I'm putting my bet on something DMA related with the controller card.

Any advice on how to run some tests on that?

I did my own little PCI bus saturation test, and I don't think it's the PCI bus at this point either.

[Vertigo Acid]

So its a no-no to use RE drives in non raid setups?It is my understanding that TLER is off by default

Someone have the data to back that up? I know i've read that here. Mage?

[unhappy_mage]

This page (http://westerndigital.com/en/products/sidebyside.asp) shows TLER as field-configurable.
in fact TLER is turned off by default with the drive, only available to toggle on with a special utility we have yet to receive from WD.

But on the WD website:
With error recovery factory set to seven seconds
I don't know what to believe. A lot of sites reference a magic utility which can enable or disable TLER, but they admit there isn't one on WD's website, and there still isn't. Anyone with a WD drive, does Data Lifeguard or one of WD's other tools change TLER?

In summary: :confused:

http://www.hardfolding.com/ftag1.php/mem/150072.png (http://www.hardfolding.com?go=38&tm=33&id=150072)http://www.hardfolding.com/utag.php/mem/1392.png (http://www.hardfolding.com?go=36&id=1392)

[hokatichenci]

Why don't you try a Linux LiveCD and see what it says with hdparm on the device? I don't have any direct access to any of my RE drives though theres nothing in the Areca configuration to change it. Either way WD probably wouldn't release a tool to change it for Linux so I'm probably SOL either way.

[John G.]

Why don't you try a Linux LiveCD and see what it says with hdparm on the device? I don't have any direct access to any of my RE drives though theres nothing in the Areca configuration to change it. Either way WD probably wouldn't release a tool to change it for Linux so I'm probably SOL either way.

No CD-ROM drive in the system. All IDE channels are being used for hard drives. :)

[hokatichenci]

No CD-ROM drive in the system. All IDE channels are being used for hard drives. :)

Try to netboot another OS then? I'm sure one of the bsds can do it. You could always yank one hard drive just to test it and then hook it back up. Open source might have a better driver implementation (assuming you're using Windows anyways).

[Justintoxicated]

I was told to get the RE drive just because it has a better warranty.

I wanted to get the 400GB RE2 but now you guys having me thinking im going to get data corruption if I do this?

[John G.]

I was told to get the RE drive just because it has a better warranty.

I wanted to get the 400GB RE2 but now you guys having me thinking im going to get data corruption if I do this?

Drive's workin' like a charm for me in a desktop environment, minus this whole PC stuttering thing (though I think it's a fair bet it's the controller card rather than the drive itself). Certainly no data corruption yet. It's only been about three days though. :)

[John G.]

Thought I'd update you guys on my situation here, in case anyone else is contemplating the use of an RE drive in a desktop environment. I've been using the drive for about six weeks now in a non-RAID environment, and this is what I've seen with my WD4000YR:

The drive stuttering I spoke of just after installing the drive ended up being the fault of the controller card entirely. Plugging the drive into another machine yields transfer rates at exactly what I'd expect, which means no stuttering. When connected to the appropriate controller card, the drive works just as you'd expect, for the most part. Nice and quiet, nice and big.

I do, however, run into the occassional problem-- especially during long-haul copy/transfer operations (as I have to fill/unload the drive every now and again, and 400GB is a lot to transfer), the drive will lock up and I'll get an I/O read error. Delayed write failure notices. Windows explorer will lock up. Basically, bad stuff. I haven't checked the integrity of every file included in 400GB, but if I do lose data, I haven't seen the evidence yet. It does scare the crap out of me though that one time I won't be so lucky.

I suppose its conceivable that this is a problem specifically with my drive, but given the "tighter tolerances" and extended warranties of these enterprise-class drives, it's probably a safe bet that TLER is responsible. Bottom line recommendation: use this drive in non-RAID environments cautiously, and if possible, for limited amounts of time. My RAID setup fell through, so now I'm looking at either bearing with this one or returning/reselling it for another. If you're going for a RAID setup, go for it from the beginning. And if you're not looking to RAID, the better warranty and higher MTBF probably aren't worth it, and I'd recommend trusting your data to an SE drive instead.

If anyone else has any experiences with RE drives in non-RAID environments, or can point me to information regarding such, please let me know.

[drizzt81]

thanks for the update :)

[unhappy_mage]

Have you checked for controller driver updates? Form the sounds of it, that'd be what I'd look at.

http://www.hardfolding.com/ftag1.php/mem/150072.png (http://www.hardfolding.com?go=38&tm=33&id=150072)http://www.hardfolding.com/utag.php/mem/1392.png (http://www.hardfolding.com?go=36&id=1392)