Re: disk problems: which ATA?
On Sun, 03 Jul 2011 14:35:08 -0700, Ross Boylan wrote:
> On Sun, 2011-07-03 at 19:19 +0000, Camaleón wrote:
>> On Sun, 03 Jul 2011 11:25:18 -0700, Ross Boylan wrote:
>> > How can I tell which ata device is which hard drive? It's come up
>> > several times for me, most recently with ata2.00: exception Emask 0x0
>> > SAct 0x0 SErr 0x0 action 0x6 frozen
>> You can:
>> - Run "smartctl -i /dev/sdb | grep -i model"
> Except the drive isn't responding to smartctl (see original message). I
> tried adding -T permissive, but all that gets me is Short INQUIRY
> response, skip product id (curiously, no error about command failed).
Yes, I already noticed but "-i" also fails? :-? It just gathers basic hdd
info, mmm... that does not sound very good.
Anyway, you can test with another hard disc utility like hdparm/sdparm
unless they're also failing.
>> - Then "dmesg | grep -i ata2"
>> - To finally compare by hdd model :-)
> That works for my current machine. But on another machine I want to
> figure out which drive an error message goes with, and there are 2
> identical drivers. I suppose that even if I new which sd device the ata
> went with, I still wouldn't be sure which physical drive that was...
Try with another tool?
>> As per the error itself, you can use the manufacturer hard disk
>> diagnostic tools which uses to run from a LiveCD and will provide
>> accurate results about your hdd health and status.
>> OTOH, I've also seen that kind of error coming from bad sata cable or
>> bad sit connection to motherboard/disk. You may also check this.
> I think I already tried reseating, but I suppose it's worth trying
> again. I'm concerned if I power down I may not be able to get back up,
> since the failing hard disk is actually part of an LVM volume group. I
> am also unable to get information on that VG right now.
What I would do is:
1/ Having a full copy of both hard disks placed into another machine,
just in case.
2/ Download manufacture's disk utility and make a full scan for both disks
> Most of the logical vomes in the group are backed by other hard drives,
> but I'm not quite sure what will happen if the disk is toast. At the
> moment, I have access to most of the LVs, even though I can't get info
> on the PV that contains them (!).
I've never used LVM so I dunno what would happen in such cases. I've
heard that it is advisable to have a RAID level on top of LVM to avoid
data loss as LVM does not prevent from this situations unless it has been
setup to run as RAID 1.
> P.S. For the record, kernel logs need to be read carefully to figure out
> which drive is ata2.
I think the important log entry was the first you provided that pointed