Disk Event ID 154

[Isgrimnur]

Well, I finally found the issue making my machine squerk.

After perusing forums, they pointed me to the Event Viewer.



Unable to locate which drive is which, I downloaded Western Digital Data LifeGuard Diagnostics. I ran the boot drive overnight on an extended test, and received no issues. I started the data drive, and it started sqwerking almost immediately.

Microsoft gives me this very informative information about doing a regedit on Windows Server 2012 machines.

This behavior can occur for one of the following reasons:

The FC connection dropped a packet somewhere between the Host Bus Adapter (HBA) and the storage array.

The array controller or a device in the array responded to an I/O request and indicated that the hardware exceeded hardware-defined time-outs in the array.

To resolve this issue, contact the vendor of your adapter, switch, or array to determine the cause of the dropped FC packets or hardware time-outs.

So it sounds like it's either an issue with the drive or the motherboard. Any thoughts or other experience with this issue?

[Rip]

I would probably try imaging the drive over to another and see if the issue remained.

Probably 75-80% the drive and eliminating that is usually cheaper/easier than doing so for the MB.

[wonderpug]

Have you tried reversing the polarity on the main reflector?

[Punisher]

try changing the cables first. probably cheapest, easiest test.

[Kasey Chang]

Only once every 7-9 seconds... That's quite a bit of time in when things are measured in GHz

[GreenGoo]

The MS blurb is hilarious, because assuming this is your own system, I have serious doubts you're running fibre channel (FC) or have any HBA's installed (think "network" cards for SAN storage).

If we can assume the Windows error is useless, which is reasonable to assume (but not always), we have almost nothing to go on. I mean, they can't even be bothered to tell you which disk is returning the io error? Bah. edit: the disk number and device name are displayed in the event viewer. What did the WD give you (besides sqawking noises I guess)?

Are you running a raid controller and if so, what raid are you using? I ask because windows desktop OS's (and older server versions too) are lazy and you will probably need to use the raid controller proprietary software to look at the controller and disks in more detail.

I don't have a lot of experience with local storage on windows but what I do know is that almost always requires 3rd party tools or, occasionally, windows plugins that aren't installed by default.

Do we know this isn't run of the mill bad sectors? I'll go back and read closer in case I missed more detail in the OP.

edit: How big is the disk? Can you just run the windows tool (scan disk? Something like that) to test the sectors? It will take forever, probably, but if it's just a few bad sectors the turning them off is an easy fix.

1. How old is the disk? Is this recent behaviour after a long period of no issues?
2. Is the disk starting to get "full"? How has this disk been utilized in the past? Lots of big data swapping, small files, slowly growing, permanent data, etc

If it's an old disk that is just starting to act up, danger will robinson (obviously). If it's a new one that has only recently started to fill up, might just be random chance upon a group of bad sectors.

I've been assuming it's an HD and not an SSD, is that correct?

edit2: Ok, found lots of others claiming that ms knowledge base article to be junk, so ignore it. Even if it identified the problem successfully the suggestion to increase your timeout setting seems seriously crazy.

Found an article that suggested using this tool: CrystalDiskInfo to check for SMART errors. Can't vouch for the tool but it sounds like a good idea.

edit3: Found another recommendation for the tool above, but also found that the WD tool you mentioned should give you the same info for WD disks (obviously).

[GreenGoo]

try changing the cables first. probably cheapest, easiest test.

This is a good idea. You can reseat the current cables first if you want. Could be a loose cable or, as Punisher suggested, a cable gone bad. You could also change SATA ports and see if the problem persists.

[Isgrimnur]

The cables were changed and moved recently. I didn't let the tool scan complete, as it was obvious that constant access resulted in repetitive sqwerks. Drive me is ~4 years old, 2TB, and mostly full. No raid. SMART shows no errors.

[GreenGoo]

Looks like WD measures reliability by life expectancy, which appears to be about 5 years for retail drives with failure rate of 0.8%.

Could just be early end of life.

Whatever the problem, backing up the data should be a priority.

edit: I should have mentioned that if the scan tool is barfing all over the place, that's probably a good sign that you're eff'd.

[Kasey Chang]

One last thing to try: SPINRITE. I know it costs $$$$, but it's a good utility to run "maintenance" on almost all hard drives. It resurrects dead drives, and keeps running drives running.

[xwraith]

This is a spinning rust drive right?

What might be happening is that the disk isn't getting a good read on the track its on so it keeps retrying, and eventually it times out the IO attempt, kicks an exception back to the OS which writes to the event log.

At any rate you probably have disk corruption. If you are hearing squeaking noises that is also a red flag to me, could be some sort of mechanical issue.

Backup your data ASAP and get ready for the disk to completely fail

[Isgrimnur]

Looks like the MB is the weakest link.

I have 6 SATA connectors on the board, 2 6GB, 4 3GB.

6G1 - main drive connection, no issues.

6G2 - drive not recognized (and I believe I moved the drive off of that one last time I was in its guts due to some other issue that I've now since forgotten)

1 - not recognized

2 - optical drive cable, disconnected, as I came home one day to it trying to spin itself up every five seconds

3 - not recognized

4 - (this year's) original position

I bought a HDD dock after last summer's issues, and had to move this drive off of 6G2.

I'm currently running the WD scanner against the drive in the dock and have heard not one sqwerk.

The board is an Asus P8Z68-V LX. Order date was 12/28/11. 0.o

[GreenGoo]

Other than trying a new cable/sata port which would have more about the adapter I never would have suspected the mb.

Well done.

[Isgrimnur]

Thanks. Once I started putting all the pieces together with last summer's issues and the optical, it really suggests the MB. I suppose I could always buy another cable to rule it out, but I feel that it's a low probability at this point.

Drive check has been running over six hours. Not a single error in the log in that time.

Now the real question is whether to try and find an LGA 1155 replacement, or move to a low end CPU/MB replacement. If I can find one that takes the same memory. Current rig is an i5-2500K @ 3.3MHz.

[Isgrimnur]

Intel BOXDH77KC on order. I'll get to spend some part of the back half of the month stripping my rig down to rebuild it.

...yay...