On Fri, Feb 6, 2009 at 16:12, Dale Ghent <[email protected]> wrote: > > On Feb 6, 2009, at 2:24 PM, Marion Hakanson wrote: >> If there's a sick/slow device, it should show up in "iostat -xn 1" with >> larger numbers in the asvc_t, wsvc_t, %w and/or %b columns. These numbers >> don't show up in the "zpool iostat" subcommands. So it does. After recreating the pool without this drive, the downspikes are gone.
> Yeah, and if you spot such a drive, follow up with a iostat -En and see if > it's reporting any errors. Check /var/adm/messages as well for any transport > errors that may be emitted by the sd driver. There are, as a matter of fact, errors reported... I think. $ iostat -En c4t9d0 c4t9d0 Soft Errors: 2 Hard Errors: 16 Transport Errors: 484 Vendor: ATA Product: SAMSUNG HD753LJ Revision: 1106 Serial No: Size: 750.16GB <750156373504 bytes> Media Error: 0 Device Not Ready: 0 No Device: 16 Recoverable: 0 Illegal Request: 2 Predictive Failure Analysis: 0 I'm guessing that the "Hard Errors" category is one to watch out for. Time to replace this disk, I guess. Would these errors eventually show up in fault management tools? Is it worthwhile to monitor the messages file for these errors? Will _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
