Re: [Linux-usb-users] Failed reads from RAID-0 array (from newbie who has read the FAQ)

["Michael Schwarz" <mschwarz@multitool.net>]

Just tried in on a stock Ubuntu Edgy install. Same thing. Locks on read.

I've got a dmesg (w/stack trace) file from the ubuntu attempt (it was
clean prior to doing the read) which I will send to Alan and Neil (any
anyone else who asks for it). There were no error messages in dmesg prior
to running the stack trace.

-- 
Michael Schwarz

> Okay. I've verified my hardware (by doing large write/reads to non-raid
> file systems on each of the seven USB flash drives on the hub).
>
> So this morning I booted cold and began gathering log data. I'm sending it
> to you guys (you hsould know this) before looking at it myself.  Here's
> the sequence:
>
> Cold boot
> Gnome login
> echo t > /proc/sysrq-trigger
> dmesg > dmesg-0-beforehub.log
> Attached hub with 7 drives
> dmesg > dmesg-1-afterhub.log
> mdadm --create /dev/md0 --auto=md --level=0 --raid-devices=7 /dev/sd?
> dmesg > dmesg-2-aftermdcreate.log
> mke2fs -b 4096 -R stride=16 /dev/md0
> dmesg > dmesg-3-aftermkfs.log
> mount /dev/md0 /mnt
> cp -rv ~mschwarz/FUTURAMA_S2D2/* /mnt
> dmesg > dmesg-4-afterbigwrite.log
> cp -rv /mnt/* fs2d2/
>
> At this point, the process hangs. So I ran:
>
> echo t > /proc/sysrq-trigger
> dmesg > dmesg-5-hungread.log
>
> ...in a different root window. All these operations were performed as root
> (in order to be as dangerous as possible -- actually, in order to reduce
> possible permissions issues; although I don't think there are any)
>
> All of these dmesg logs are attached in gzip format. I don't know what
> majordomo will do with those, but the cc's going directly to Alan and Neil
> should come through. I'm going to start combing these files myself, so if
> you guys want to save time, you can certainly give me a couple hours to
> get started! ;-)
>
> As always, I'm very grateful for your assistance and that of the group!
>
> --
> Michael Schwarz
>
>> Yeah, I understand that.
>>
>> Sorry, I use squirrelmail. Pretty limited...
>>
>> I'll get you a "raw" dmseg output when I replicate the problem.
>>
>