On 7/14/07, Lan Barnes <[EMAIL PROTECTED]> wrote:
On Sat, July 14, 2007 8:18 am, Mark Schoonover wrote:
> On 7/13/07, James G. Sack (jim) <[EMAIL PROTECTED]> wrote:
>>
>> Mark Schoonover wrote:
>> > On 7/13/07, James G. Sack (jim) <[EMAIL PROTECTED]> wrote:
>> >>
>> >> Lan Barnes wrote:
>> >> > On Fri, July 13, 2007 7:02 pm, James G. Sack (jim) wrote:
>> >> >> Lan Barnes wrote:
>> >> >>> My MythTV box (FC6) is throwing these messages on boot and in
>> dmesg:
>> >> >>>
>> >> >>> hda: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
>> >> >>> hda: drive_cmd: error=0x04 { DriveStatusError }
>> >> >>> ide: failed opcode was: 0xb0
>> >> >>> hdb: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
>> >> >>> hdb: drive_cmd: error=0x04 { DriveStatusError }
>> >> >>> ide: failed opcode was: 0xb0
>> >> >>> hdd: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
>> >> >>> hdd: drive_cmd: error=0x04 { DriveStatusError }
>> >> >>> ide: failed opcode was: 0xb0
>> >> >>>
>> >> >>> Both drives are less than a year old; one less than 6 months.
>> What's
>> >> >>> going
>> >> >>> on?
>> >> >>>
>> >> >> I've seen these error messages for
>> >> >> 1) actual bad drives (but usu. not where 2+ fail at once)
>> >> >> (confirmed by manufacturer's diagnostic utility)
>> >> >> (preferably, testing on another box)
>> >> >> 2) bad cabling or loose connectors
>> >> >> 3) bad motherboard (I think) -- or maybe heat problem in
chipset?
>> >> >>
>> >> >> In some cases I have regained stability by backing off the UDMA
>> >> settings
>> >> >> via hdparm -X until the DriveReady SeekComplete Error stopped.
>> Can't
>> >> >> say I felt good (or comfortable) about doing that, though.
>> >> >>
>> >> >> Regards,
>> >> >> ..jim
>> >> >>
>> >> >
>> >> > The box appears to be running beautifully. How afraid should I be?
>> >> >
>> >>
>> >> My experience is that this kind of error message seemed to get worse
>> >> over time (longer streams of messages, etc), and that eventually
>> showed
>> >> up as sluggish behavior and ultimately, hard 'lockups'.
>> >>
>> >> Disclaimer: I'm speaking with no hardware credentials here -- just
my
>> >> personal amateurish assessment.
>> >>
>> >> Perhaps someone more knowledgeable will comment on the particular
>> error
>> >> messages you quoted?
>> >>
>> >> Regards,
>> >> ..jim
>> >
>> >
>> >
>> > When I see these error messages in my server logs, the drive gets
>> replaced.
>> > Drives are too cheap to risk data loss.
>> >
>>
>> Note that he's logging errors for hda, hdb, and hdd (one of which is
>> probably a CD) -- do you see anything suspicious in that?
>>
>> Regards,
>> ..jim
>>
>
> Not really actually. In my case, I got the same errors on a Gentoo
system.
> Drives were the same, hda, hdb and hdd. Hdd was the CD in that system.
Hda
> and hdb were mirrored, but on separete controllers, with hdd being on
the
> same controller as hda. I replaced hdb because it was the easiest drive
to
> reach. Odd thing was, hdd stopped generating errors. Hda was still
> throwing
> these errors, so I replace that too. Md handled things just fine, so no
> data
> loss.
>
> It's the most puzzling thing I've seen. I was thinking like you, it's
mobo
> issues, PS issues, something central to the system based on all the
drives
> throwing these errors. I just ended up with two drives going bad, and
yes
> they were from the same manufacturing batch.
> --
Are the drives really bad, or is smartd bad? I used smartdctrl to spill
the data, and it claims they're old old old. No way.
I have heard such nasty things about smartd that I'm tempted to archive
the only really valuable stuff (family photos) to DVD, which I should do
anyway, turn off smartd, and soldier on.
Comments?
--
Lan Barnes
SCM Analyst Linux Guy
Tcl/Tk Enthusiast Biodiesel Brewer
Well, smartd saved my bacon. It was correct, the drives were going bad. For
IDE drives, I rely on Spinrite to test them out. It'll do things to your
drives the manufacturer never thought of, plus it can try to recover bad
blocks. I burn in all my new drives with Spinrite before they go into a
server. It's worth the couple of days burning the drives, I've found a
couple brand new drives that had too many bad blocks, so I returned them.
Archiving to CDR/DVD is also problematic. Where I work, we archived data to
CDR about 4 years ago, and on some of these disks, we couldn't read them any
longer. They were stored in a zipped up holder, inside a library like room.
No sunlight could get to them. We ended up copying ~450 CDRs back to drives.
--
Mark Schoonover, CMDBA
[EMAIL PROTECTED]
http://marksitblog.blogspot.com
Cell: 619-368-0099
Database Administration * System Engineering * Software Development *
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
