PDA

View Full Version : Issue with new 2.3GHz G5 Tower Hard Drives?




Dave Marsh
Jun 2, 2005, 01:04 AM
Hi all. I was setting up a new dual 2.3GHz Tower today with 1GB memory, a Sony SuperDrive, and a Western Digital 250GB hard drive. After transferring the user stuff to the new Mac with the Migration Assistant, and running Software Update to bring everything up to spec, I ran Disk Utility to Repair Permissions (all OK), Disk Warrior 3.0.3 to check the file structure (all OK), and iDefrag 1.1b1 from Coriolis-Systems (a new app to replace Norton's venerable Speed Disk).

About 50 minutes into the defrag, iDefrag stalled. I was running the Mac from my FireLite FireWire drive (running MacOS X 10.4.1), and everything else seemed to be fine, so I was a bit puzzled. I've been using iDefrag for a couple of months now, and it has worked reliably for me. I entered into an email exchange with their support staff and they were puzzled as well. They suggested that I track issues being recorded in the system.log, as well as their own log, so I did.

It turns out that whenever iDefrag stalled, a few seconds later the system.log recorded this entry: kernel[0]: IOATAController device blocking bus error, and proceeded to continue to post a new entry every few seconds until I rebooted the Mac.

I had initially thought there might be a bad block on the hard drive, and had run a lengthy surface scan on the platters using Tech Tool Pro 4.0.4, but nothing turned up. A check of the volume's directory and file structure also reported no errors. So, I proceeded to repeatedly reboot and continue with iDefrag (three times as it turns out) since it was making progress (i.e., not hanging at the same place on the platter). iDefrag eventually successfully completed the defrag with no errors, so I'm now very suspicious about the hard drive.

This evening Coriolis Systems got back to me to say they saw some hard drive issues being reported with this log entry from last year, and sure enough there are. Those postings (search on IOATAController in Google) mostly applied to Mac laptops, but some users were reporting the issue with Western Digital drives, as well.

I'm now suspecting that I may have a hard drive issue. I left the Mac running overnight and look forward to checking on its status in the morning. The issue manifested itself every 45-50 minutes following intense disk read-write activity, which it won't get tonight.

However, I thought I'd post this online to alert others to a potential issue.

More to follow later. :confused:



Dave Marsh
Jun 2, 2005, 03:33 PM
Apple agreed today to ship me a replacement drive to resolve this issue. I hope the replacement doesn't have the same issue!

gutilin
Jun 17, 2005, 06:16 AM
Im a OSX system administrator. we have mostly Power Macs G4 1Ghz and today they called me because they couldnt start the computer. From the very beginning it looked like a hard disk problem because if you tried to reinstall OSX it wont detect the hard disk. This theory seemed even more true when you realized that the computer sometimes found the HD and sometimes not. And the startup was interrupted at different steps. The strange thing was that you could mount the computer from another via Firewire and it worked OK.

Right now I could make it start in Safe mode (pressing shift at the start-up). It took long time but once the system started everything was working OK. The first thing I did was checking the warning console and the most eyecatching error for me was the following:

IOATAController device blocking bus

So right now, and before even trying anything else, Im backing up 22Gb of the user data to another disk. Ill tell you if it finally works.

Im suspicious about being a HD hardware problem but I read somewhere that someone called to Apple and they told him it seemed to be a "controller loosing logic" (????).

What is a matter of fact is that the darwin IOATAController file is used to develop hard disks drives. But if my big backup works I might discard being a HD problem and I might think it to be a kernel extension problem.

Any help will be strongly appreciated!!

Dave Marsh
Jun 17, 2005, 09:52 AM
Apple sent me a replacement 250GB SATA-HD (Maxtor) overnight to replace the problematic Western Digital drive. I plugged it in, and used SuperDuper! to clone it to the new drive. The replacement drive works flawlessly, and the issues that the new WD drive was exhibiting with those error messages in the system.log have stopped. That's why I don't think it was a software issue. The data/system/apps on the replacement drive are IDENTICAL to the original drive with the issue. :D

Dave Marsh
Nov 28, 2005, 07:36 PM
As luck would have it, the 250GB hard drive in my new personal 2.7GHz G5 Tower is also a Western Digital and I had this same issue crop up this past month. Apple's position this time is that since iDefrag is the only application that appears to have a problem with this drive, the problem must be Coriolis Systems' issue to resolve. In that regard, Coriolis Systems has opened a ticket with Apple, and they have also brought the issue up with their Apple Developers contacts, and with Western Digital itself.

For Apple's part their Help Desk supervisor called me this evening to notify me that I should either switch to Tech Tool Pro's defragmenter, or purchase a different vendor's hard drive to use, and that he was marking my Help Desk ticket as "Resolved." :mad:

With nowhere left to turn, I've begun researching hard drives and will probably be replacing my startup drive with a Hitachi 250GB T7K250 SATA II drive. The offending controller in my G5 Tower is only SATA I, but the Hitachi SATA II drive is supposed to be fully backwards compatible and costs virtually the same as the old SATA I model, so I'll probably be able to use it later if/when I replace my current tower.

If anything new arises, I post an update, but as of this writing Apple has disappointed me greatly. I doubt I'll be quite the advocate for Apple's support philosophy in the future as I have been over the past 20 years.:(

Bear
Nov 28, 2005, 08:16 PM
...
With nowhere left to turn, I've begun researching hard drives and will probably be replacing my startup drive with a Hitachi 250GB 7T250 SATA II drive. The offending controller in my G5 Tower is only SATA I, but the Hitachi SATA II drive is supposed to be fully backwards compatible and costs virtually the same as the old SATA I model, so I'll probably be able to use it later if/when I replace my current tower.
...Apple ships the Hitachi 7k500 SATA II drive in the PowerMacs when you CTO 500GB drives. So the 250GB SATA II should work just as well.

As for Western Digital, I would never willingly buy one of their drives.

Dave Marsh
Nov 28, 2005, 08:28 PM
Thanks for the follow-up. I'm looking forward to getting the Hitachi drive now. It also appears to be a bit faster than the WD.

Monyx
Nov 30, 2005, 05:23 AM
I'm curious why do you defrag under OSX when Apple advise that there is "little benefit to defragmenting"?

http://docs.info.apple.com/article.html?artnum=25668

Dave Marsh
Dec 1, 2005, 03:41 PM
Two reasons:
1. MacOS X only automatically defrags files <20MB
2. iDefrag also REARRANGES the files on the drive to place those most often used closer together for more rapid access.

Yes, today's drives are pretty fast. But, I've observed noticeable performance improvements (items opening faster) after doing a defrag. Once every couple of months seems to do the trick.

jkraus
Feb 15, 2006, 12:37 PM
It turns out that whenever iDefrag stalled, a few seconds later the system.log recorded this entry: kernel[0]: IOATAController device blocking bus error, and proceeded to continue to post a new entry every few seconds until I rebooted the Mac.


Same problem here with WDC hard drive. Apple care did not acknowledge the incompatiblity between 3rd party software and WDC hd. So no new hard drive in my case.

On top of all this, while recovering from the problem, new ones arose. Maybe one could help me here:

System profiler of iMac G5 fails after timeout of the iDefrag defragmentation program...only works now from the external hard drive. DiskWarrior, fixing prefs and running cron scripts did not help. I tried to retrieve the system profile from the terminal and get the paragraphs below.

Question: what causes this, what does the error mean, and how can it be fixed?

It has been established in the meantime, that the Western Digital hd causes a timeout in the defragmentation program...which results in a freeze.
It turns out that whenever iDefrag stalled, a few seconds later the system.log recorded this entry: kernel[0]: IOATAController device blocking bus error, and proceeded to continue to post a new entry every few seconds until I rebooted the Mac.

Apples does not acknowledge this imcompatiblity between third party software and Western Digital drive.

Here the system-profiler related error message


Last login: Tue Feb 14 23:59:01 on ttyp1
Welcome to Darwin!
jurgen-kraus-imac-g5:~ jkraus$ /usr/sbin/system_profiler
2006-02-15 00:02:17.732 system_profiler[293] CFLog (0):
CFPropertyListCreateFromXMLData(): plist parse failed; the data is not proper UTF-8. The file name for this data could be:
/System/Library/Frameworks/Foundation.framework/Resources/Languages/en
The parser will retry as in 10.2, but the problem should be corrected in the plist.
2006-02-15 00:02:17.797 system_profiler[293] CFLog (0):
CFPropertyListCreateFromXMLData(): plist parse failed; the data is not proper UTF-8. The file name for this data could be:
/System/Library/Frameworks/Foundation.framework/Resources/Languages/en
The parser will retry as in 10.2, but the problem should be corrected in the plist.
2006-02-15 00:02:17.828 system_profiler[293] CFLog (0):
CFPropertyListCreateFromXMLData(): plist parse failed; the data is not proper UTF-8. The file name for this data could be:
/System/Library/Frameworks/Foundation.framework/Resources/Languages/en
The parser will retry as in 10.2, but the problem should be corrected in the plist.
2006-02-15 00:02:17.879 system_profiler[293] CFLog (0):
CFPropertyListCreateFromXMLData(): plist parse failed; the data is not proper UTF-8. The file name for this data could be:
/System/Library/Frameworks/Foundation.framework/Resources/Languages/en
The parser will retry as in 10.2, but the problem should be corrected in the plist.
2006-02-15 00:02:17.972 system_profiler[293] CFLog (0):
CFPropertyListCreateFromXMLData(): plist parse failed; the data is not proper UTF-8. The file name for this data could be:
/System/Library/Frameworks/Foundation.framework/Resources/Languages/en
The parser will retry as in 10.2, but the problem should be corrected in the plist.
2006-02-15 00:02:18.031 system_profiler[293] *** -[NSBundle load]: Error loading code /System/Library/SystemProfiler/SPSerialATAReporter.spreporter/Contents/MacOS/SPSerialATAReporter for bundle /System/Library/SystemProfiler/SPSerialATAReporter.spreporter, error code 2 (link edit error code 0, error number 0 ())
2006-02-15 00:02:18.031 system_profiler[293] *** Uncaught exception: <NSInvalidArgumentException> *** -[NSPlaceholderString initWithString:]: nil argument
Trace/BPT trap

Dave Marsh
Feb 15, 2006, 01:31 PM
Sorry to hear about your problem, but it confirms the compatibility issue with the Apple I/O controller and the Western Digital 250GB hard drive controller.

I'm reading that your System Profiler no longer works properly and your iMac is now sluggish and behaving abnormally. Did Disk Warrior complete successfully, and fix all issues it found? If so, that's good, even through your Mac is still sick. Disk Warrior is primarily concerned with assuring that files/fragments identified in your hard drive's catalog are where it says they are on your hard drive. If it's confirmed that, and after Disk Utility has VERIFIED your hard drive's structure is OK, the problem must be elsewhere.

At this point, I'd have to guess that a properly identified file has been corrupted. Make sure Journaling is enabled on your drive when you run Disk Utility.

With Journaling enabled, Disk Warrior reporting all's OK, Disk Utility has checked permissions and verified that the hard drive structure is OK, the next step is to rerun your latest MacOS X updater.

If you're running MacOS X 10.4.4, I'd download and run the 100MB+ combined updater over. That will compare the system files currently installed and replace those that are different. MacOS X 10.4.5 was released yesterday, so I'd run that as well after a successful update with 10.4.4. If you can find the 10.4.5 combined updater (I don't know if it's out yet), I'd use it instead, but be sure you run the combined updater (>100MB), not the patch (~6MB).

If those file replacements run OK, it will take a while for MacOS X to restart all the way to the Desktop. Let it run... It may take more than 100 spins of the startup wheel, and then an equal amount of blue screen while it rebuilds the Desktop environment. Be patient.

If, after 20 minutes or so (it depends on number of changes it made, memory available, and speed of your CPU) it hasn't finished, you may have to force quit (ungraceful shutdown) and reinstall from your DVD installer (using Archive & Install), but I'd NOT do this until I'm convinced it has stopped updating (listen for hard disk activity). If the hard drive is chugging away, leave it alone. If it seems to be stuck (repetitive hard drive activity that repeats the same over and over), you may be out of luck and have to rebuild the OS.

The Western Digital hard drive seems to work OK when it's not stressed with intensive read/write activity for more than 10 minutes or so. Coriolis-systems has a ticket opened with them, and with Apple, but I haven't heard from them in awhile. This means to you that you should be able to get your drive working again successfully without replacing it. You just won't be able to use iDefrag on it.

I resolved my issue by purchasing an internal Hitachi 250GB HD, using SuperDuper! to clone my Western Digital's content over to it, and resetting my Hitachi as my startup drive. I reformatted the WD drive, and now let SuperDuper! do a nightly backup to the drive from the Hitachi. The WD drive controller doesn't seem to mind intensive write activity, just intensive read/write activity, so the nightly backups are working great.

I'll let Coriolis-systems know another Apple customer has been bitten by this glitch and point them to this forum for additional ammo in their quest to resolve the issue with WD and Apple.

Good luck. Just be assured that the WD HD's hardware seems to be OK. The issue appears to be I/O controller related and can be avoided by not attempting intensive (over 10 minutes) read/write activity on it. Apple's being wrong-headed about not swapping out the drive for those customers having this issue who simply want to keep their drives in top shape, but there doesn't appear to be anyway to force them to do the right thing. Your system CAN be fixed, but it may take awhile for you to get everything back up to snuff.

jkraus
Feb 19, 2006, 11:23 PM
[QUOTE=Dave Marsh]

I'm reading that your System Profiler no longer works properly
Yes!

and your iMac is now sluggish and behaving abnormally.
No! Appears fine.

Did Disk Warrior complete successfully, and fix all issues it found?
Yes!



At this point, I'd have to guess that a properly identified file has been corrupted.
iDefrag's opinion: It looks like there's a corrupt property list file on your disk. You could
probably find out which file it was by using fs_usage, but who knows what other
damage you might have. It's probably safer to just reinstall.



With Journaling enabled, Disk Warrior reporting all's OK, Disk Utility has checked permissions and verified that the hard drive structure is OK, the next step is to rerun your latest MacOS X updater.

Will do!


I resolved my issue by purchasing an internal Hitachi 250GB HD,

What about Seagates SATA II...that's waht's presently on offer here. I am afraid the Seagates run hotter than the WDCs.


I am also not happy with Apple's argument that it is the third party's software's fault. I purchased DriveGenius for a disk surface scan. It also contains a defragmentation program, but that is not a thorough as iDefrag. I will not use Techtool Pro or anything else by Micromat, as their Drive X in 2001 never fixed anything hwoever totally corrupted the volume structure.

Thanks,
Jurgen

OK, I called Apple again: this time, they were willing to replace the WDC hd with another WDC hd...they claim that, in Canada, they can only replace part with such from the same manufacturer. Strangely enough, my faulty Matshita superdrive was replaced with a Pioneer one.

Will stay at it.

Jurgen

PS. No wonder Europeans are being considered arrogant.

1. I fixed permissions and ran cron scripts
2. I ran diskwarrior
3. I started the machine in safe mode
4. I ran the combined os 10.4.5 update combo
5. I reset the SMU


I did not help. Apple sent me a capture utility which records all logs and system information. It went into an endless loop whenever it arrived at the system profile.


I re-loaded the os...whatta pain.

In the meantime, I got Apple to look into the I/O controller bus blockage error. I also put Coriolis software in contact with Apple.

case # 59385507



Thanks again.
Jurgen

I am quoting a local guy who tried to help...good idea:

Interesting. Having developed a lot of low level driver software including disk drivers. I would be that the WD drives have a setting that needs to be set/cleared that they are not doing themselves and Apple is not doing in their drivers. As they write their own code on the Intel box it would be believable that they make the correction, but because Apple writes their own code to the specs they would not know about it.

I saw a similar problem on the early MO drives (I helped pioneer them),

Regards,
Ron

alexprice
Feb 20, 2006, 05:48 PM
Hmmm, so let's say that you think you have found an issue with the drive, what then, a backup, buy new drive? hard drive's always die in the end. If you have backups then no need to worry, 50% of all Mac repairs I perform are faults with hard drives made by Fujitsu, Maxtor, Hitachi, Seagate etc. We cannot blame Apple either as they didn't make the drives, they cannot work out in advance if they are going to be given a good batch from their suppliers. As for this defragging program, Surely the built-in defragging does the job.

If it works, don't fix it.

jkraus
Feb 21, 2006, 03:27 PM
Thanks Alex,
Yes, drive as such is ok, so why should I buy a new one? Sure, it is not Apple's fault when a drive does not work, however it is Apple's responsibility to make sure the components they use in their machine are compatible. My problem is likely a firmwire issue and yes, Apple are eventually responsible for it. They should replace the drive and then dedice whether to go after WDC or not.

In terms of built-in defragging: it does not do thge job as it ignores files above 20 MB and also free space. Even Prosofts Drive Genius does a rather sloppy job.

Jurgen:D