On Fri, 11 May 2007, Sophia Li wrote:
> >>>> -------- Original Message --------
> >>>>On 5/10/07, Al Hopper <[EMAIL PROTECTED]> wrote:
> >>>>
> >>>>> My personal opinion is that USB is not robust enough under (Open)Solaris
> >>>>> to provide the reliability that someone considering ZFS is looking for.
> >>>>> I base this on experience with two 7 port powered USB hubs, each with 4
> >>>>> *
> >>>>> 2Gb Kingston flash drives, connected via 2 ports to a Solaris (update 3)
> >>>>> desktop box which runs ZFS on two internal 500Gb drives. I see about 24
> >>>>> to 28Mb/Sec (bytes) maximum of bandwidth over each USB bus. One time,
> >>>>> after disconnecting one hub (to show someone the hub with 4*USB drives)
> >>>>> it
> >>>>> hung the OS and reset the box. A subsequent import of the ZFS volume
> >>>>> that
> >>>>> was disconnected, failed. (Yes it was exported, but failed to import).
> >>>>> So my take on USB is ... it's not sufficiently robust - and a USB
> >>>>> related
> >>>>> failure is likely to cause loss of the entire ZFS dataset; i.e., its
> >>>>> likely to trash more that one drive in a raidz config.
>
> I am interested in your this comment on USB. But it seems too general
> and not helpful to solve problems. Several issues have been mixed
> together which may not necessarily be USB's fault. If you believe there
Agreed.
> is a USB issue, a better practice is to file a bug. And please make sure
> the problem is reproducible and be detailed in problem description. :-)
Understood.
> I play with USB devices a lot and seldom see hotplugging hangs a system.
> The hang looks very exceptional to me. Could you experiment more with
> the devices and combinations of filesystem configuration? e.g., if you
> put on the drives UFS instead of ZFS, would it hang? Is there a way that
> you can reproduce the hang much more reliably?
I really can't experiment with this particular machine, because its my
main desktop that drives a 22" and 30" LCDs, has about 18 Gnome workspaces
and 80+ windows active. If the Xserver dies it take 10 to 15 minutes just
to get everything setup again so that I can get my productive development
environment in place. And that is aside from the ZFS mirrored pool on the
machine that has 45+ filesystems and 6 zones defined. So - experimenting
with it is not possible for now. ... more below ...
> Another question is if you are using ZFS on USB drives, the system hangs
> due to non-usb related reason and you reset the box, can data integrity
> on the USB drives be ensured?
>
> Yet another question is if you are using non-USB drives, the system
> hangs due to whatever reason and you reset the box, can data integrity
> on the non-USB drives be ensured? And how, by SW or HW?
>
> We need to think of the questions and make clear if such kind of data
> loss is particular to USB or not before coming to a conclusion too quickly.
The only conclusion I've reached is that attaching 6 or 8 750Gb disk
drives via USB for use as ZFS pool is not a good idea - because USB is not
robust enough to guarantee, that in the event of a USB failure or "event",
that no more than one disk configured in a raidz configuration will be
negatively impacted. If more than one disk drive in a raidz storage
pool is negatively impacted, then you'll loose the entire pool. USB is
not an appropriate bus to support that type of usage scenario IMHO.
USB is fine as a demonstrator of ZFS capabilities by connecting multiple
USB flash drives (as I've done) and can also be used as a way to archive
files reliably on a removable media (the flash drives). And using ZFS
with flash drives solves the problem of corrupted or bad sectors on
low-cost flash drives - which may or may not have been 100% tested before
they were sold.
>
> >>>>
> >>>>> I'd be interested in hearing other opinions on USB connected drives
> >>>>> under (Open)Solaris ....
> >>>>
>
> Any bus can have errors. USB is nothing particular, just the chance of
> encountering errors is bigger since USB device is cheap. But isn't the
Every bus topology has appropriate and inappropriate uses - and USB is no
exception to that rule.
> file system expected to handle possible errors?
see above.
Regards,
Al Hopper Logical Approach Inc, Plano, TX. [EMAIL PROTECTED]
Voice: 972.379.2133 Fax: 972.379.2134 Timezone: US CDT
OpenSolaris Governing Board (OGB) Member - Apr 2005 to Mar 2007
http://www.opensolaris.org/os/community/ogb/ogb_2005-2007/
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
