Re: FLUSH CACHE EXT timeouts with ST31500341AS

["Anssi Hannula" <anssi.hannula@iki.fi>]

Anssi Hannula kirjoitti:
> I'm continously getting the following error with a Seagate ST31500341AS
> when under load:
>
> ata17.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
> ata17.00: failed command: FLUSH CACHE EXT
> ata17.00: cmd ea/00:00:00:00:00/00:00:00:00:00/a0 tag 0
>          res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x4 (timeout)
> ata17.00: status: { DRDY }
> ata17: hard resetting link
> ata17: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
> ata17.00: configured for UDMA/100
> ata17.00: retrying FLUSH 0xea Emask 0x4
> ata17.00: device reported invalid CHS sector 0
> ata17: EH complete
>
> Reading [1] it seems that this is a known issue with this firmware when
> NCQ is enabled, and kernel disables NCQ automatically due to this.
> However, I'm experiencing the issue despite this. This is on kernel
> 2.6.33.1.
>
> NCQ seems to be correctly disabled:
> # cat /sys/block/sdo/device/queue_depth
> 1
>
> It seems likely to me that the firmware issue is therefore not limited to
> NCQ. Or do you think this might be caused by something else? Do you need
> any more information?
>
> In the meantime, I'll try to contact Seagate support to see if they have a
> firmware update for my drive.

Indeed they have one (their serial number search didn't work for my drive
so I didnt notice it earlier):
http://seagate.custkb.com/seagate/crm/selfservice/search.jsp?DocId=207957

I'll be updating my firmware soon, but if needed I can do some testing
with the old firmware before that. If the update helps, I guess the wiki
should be updated to say that disabling NCQ doesn't completely fix the
issue.

> [1]
> https://ata.wiki.kernel.org/index.php/Known_issues#Seagate_harddrives_which_time_out_FLUSH_CACHE_when_NCQ_is_being_used

-- 
Anssi Hannula