Device Removed by Administrator in ZPOOL?

[Karl Denninger]

On 4/5/2014 8:45 PM, Kaya Saman wrote:
> On 04/06/2014 01:28 AM, kpneal at pobox.com wrote:
>> On Sun, Apr 06, 2014 at 01:12:33AM +0100, Kaya Saman wrote:
>>> Many thanks for the response!
>>>
>>> The server doesn't show any lights for "drive error" however, the blue
>>> read LED isn't coming on, on the drive in question (as removed from 
>>> ZPOOL).
>>>
>>> I will have a look for LSI tools in @Ports and also see if the BIOS LSI
>>> hook comes up with anything.
>> Have you seen any other errors in your logs? Seems like if a drive fails
>> there should be some other error message reporting the errors that 
>> resulted
>> in ZFS marking the drive removed. What does 'dmesg' have to say?
>>
>> Once ZFS has stopped using the drive (for whatever reason) I wouldn't
>> expect you to see anything else happening on the drive. So the light not
>> coming on doesn't really tell us anything new.
>>
>> Also, aren't 'green' drives the kind that spin down and then have to 
>> spin
>> back up when a request comes in? I don't know what happens if a drive 
>> takes
>> "too long" to respond because it has spun down. I have no idea how 
>> FreeBSD
>> handles that, and I also don't know if ZFS adds anything to the 
>> equation.
>> Hopefully someone else here will clue me/us in.
>
> Ok this is really weird.... just did a reboot and now:
>
> $ zpool status
>   pool: ZPOOL_2
>  state: ONLINE
> status: One or more devices is currently being resilvered.  The pool will
>     continue to function, possibly in a degraded state.
> action: Wait for the resilver to complete.
>   scan: resilver in progress since Sun Apr  6 02:43:03 2014
>         1.13G scanned out of 7.77T at 22.2M/s, 101h57m to go
>         227M resilvered, 0.01% done
> config:
>
>     NAME        STATE     READ WRITE CKSUM
>     ZPOOL_2     ONLINE       0     0     0
>       raidz2-0  ONLINE       0     0     0
>         da0     ONLINE       0     0     0
>         da1     ONLINE       0     0     0  (resilvering)
>         da2     ONLINE       0     0     0
>         da3     ONLINE       0     0     0
>         da4     ONLINE       0     0     0
>
>
> ???? Looks like the drive might have fallen off the controller?
>
> Am just looking at the tools for it on the LSI website but there 
> doesn't seem to be anything FreeBSD related.... Linux and Solaris yes 
> but no FBSD?
>
> Model is LSI SAS 9207-4i4e
>
It looks like the drive detached itself.  I've seen those "Green" drives 
do this before; they go to "sleep" if quiescent, and sometimes fail to 
wake up properly.  The controller then detaches them thinking they're 
dead, but they're not...

I'd get those things off your system.  They work ok for desktop PCs but 
I don't like them in servers.

-- 
-- Karl
karl at denninger.net


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2711 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20140406/2b94f6f1/attachment.bin>