Troubleshooting intermittent stall of one WD15EADS-00P8B0

[fgrieu]

Hello,

I have a problematic WD15EADS-00P8B0 drive. On two different machines, several times per hour of use, it 'stalls' for like a minute, sometime more, with the drive led activity lit fixed, no seek noise, and nothing hapenning. Waiting always is a cure. No data was ever lost after dozens of such events. Windows logs show nothing. The symptom occured shortly after a fresh install.

Various tests I have run pass (DLGDIAG5 for DOS, DLGDIAG for Windows, MHDD surface scan).

SMART infos (read with something that actually reports them) show no obvious problem: the drive has no reallocated sectors, 457 hours of power-on time, 125 power-on cycles, 8162 head load/unload.

I did clean installs of XP on a trusted machine from original XP Pro SP3 media, no net connection, and got that same symptom with that drive before installing anything not on that CD.

I tried a clean install on another WD15EADS-00P8B0 (same machine, drive location, cables, media..), and did not observe the problem, thus I believe that the fault is with the drive -but can't be sure, much less demonstrate it-.

What should I do next ? Short of hoping the drive fails with a more clear symtom so that I can RMA...

[Nathan_H]

That does sound suspicious.  When you tried with a second computer, did the that computer have a different motherboard or SATA controller chipset?

If your gut feeling is that the drive is bad, I would definitely exchange it at your place of purchase.  If you're outside of the return period, go ahead and create an RMA through Western Digital.  Your drive shouldn't be going unresponsive like that.

[fgrieu]

Update: the issue worsened and the drive now fails its short self test, by DLGDIAG5, which records the error in its log. Got an RMA #

---

I have seen the problem on two different machines:

- one Asus with Amd 760G chipset

- one Gigabyte with intel ICH10 SB.

That has been with 2 installs of XP on the first machine and 3 now 5 on the second. 

SMART infos are OK. All tests of DLGDIAG5 pass.

On one attempt I did a full zero of the drive (overnight), then clean XP install, no network connection, not even the graphic drivers. The problem at first did not occur. But the next morning shortly after boot, with the install still virgin, got something like a 2 or 3 minutes stall with the hard disk light lit fixed, upon pulling the Start menu. This time I put my ear next to the drive and I think it was seeking repeatedly.

I have another WD15EADS-00P8B0 with the same FW (01.00A01) and build date (03 SEP 2009), and several WD10EADS. I have not seen any problem with them but the usage pattern has been very different (backup drives). I have freed the WD15EADS-00P8B0 to make one clean XP test. I have seen no problem.

I have no precise theory so far. The problem could be related to the head unload timer (as far as I can tell, the problem does not seem to occur during period of heavy drive activity). It could appears to be be a bad drive. Or it could be that the drive internally use 4kiB sectors and sometime must make complex things to empty its write cache due to alignment issues: I use the XP CD to format, which starts the partition at sector 63, but my understanding is that WD claims that should be immaterial. Perhaps I should retry with an aligned partition..I tried to align the partition to a 4kiB boundary with linux's fdisk, no change. At that stage I'm very motivated to dertermine EXACTLY what the problem is.

In one recent event the stall occured while I was keying-in the first account name during the XP installation procedure. There was some head-moving-like noise with a long (>1 second) repeat pattern. The problem occured after I made a pause of like a minute, which likely was enough for the drive to unload its head. This might be part of the mechanism that trigger the delay. [WIndows does not put the drive to sleep: that's not enabled by default].

Secure erase of the drive (the closest thing I know to physical format) does not cure the issue (under MHDD: PWD then FASTERASE then WAIT; lasts about 5 hours and a half; FWIW, I got 5h 28' 13" and 5h 28' 12" on two tests with the "bad" drive, and 5h 05' 36" or less on two tests with the "good" drive).

[fgrieu]

In one recent event the stall occured while I was keying-in the first account name during the XP installation procedure. There was some head-moving-like noise with a long (>1 second) repeat pattern. The problem occured after I made a pause of like a minute, which likely was enough for the drive to unload its head. This might be part of the mechanism that trigger the delay. [WIndows does not put the drive to sleep: that's not enabled by default].

Secure erase of the drive (the closest thing I know to physical format) does not cure the issue (under MHDD: PWD then FASTERASE then WAIT; lasts about 5 hours and a half; FWIW, I got 5h 28' 13" and 5h 28' 12" on two tests with the "bad" drive, and 5h 05' 36" or less on two tests with the "good" drive).