[aklug] Re: Sudden slow file system access

From: Christopher E. Brown <cbrown@woods.net>
Date: Wed Jan 19 2011 - 21:38:52 AKST

You have failing SATA interface, loose cabling (drive power or SATA), or
most likely a failing HD.

If you don't already have, make full backups now and prepare for a restore
on a new drive.

On Wed, 19 Jan 2011, Christopher Howard wrote:

> Suddenly I am having a problem with my Linux desktop PC. Applications
> and desktop are suddenly becoming super slow and lurching. It seem to be
> related to the filesystem; i.e., whenever something tries to access any
> of my filesystems, response back is super slow, unless the data was
> recently accessed (cached).
>
> Syslog keeps giving me output like this over and over again:
>
> Jan 19 21:10:43 enigma kernel: [ 3068.162953] ata2.00: exception Emask
> 0x10 SAct 0x0 SErr 0x1810000 action 0xe frozen
> Jan 19 21:10:43 enigma kernel: [ 3068.162960] ata2: SError: { PHYRdyChg
> LinkSeq TrStaTrns }
> Jan 19 21:10:43 enigma kernel: [ 3068.162964] ata2.00: failed command:
> FLUSH CACHE EXT
> Jan 19 21:10:43 enigma kernel: [ 3068.162970] ata2.00: cmd
> ea/00:00:00:00:00/00:00:00:00:00/a0 tag 0
> Jan 19 21:10:43 enigma kernel: [ 3068.162972] res
> d0/00:00:4c:33:e0/00:f0:b8:27:e0/a0 Emask 0x12 (ATA bus error)
> Jan 19 21:10:43 enigma kernel: [ 3068.162975] ata2.00: status: { Busy }
> Jan 19 21:10:43 enigma kernel: [ 3068.162981] ata2: hard resetting link
> Jan 19 21:10:43 enigma kernel: [ 3068.162984] ata2: nv: skipping
> hardreset on occupied port
> Jan 19 21:10:49 enigma kernel: [ 3074.083019] ata2: link is slow to
> respond, please be patient (ready=0)
> Jan 19 21:10:50 enigma kernel: [ 3074.644035] ata2: SATA link up 3.0
> Gbps (SStatus 123 SControl 300)
> Jan 19 21:10:50 enigma kernel: [ 3074.669320] ata2.00: configured for
> UDMA/33
> Jan 19 21:10:50 enigma kernel: [ 3074.669326] ata2.00: retrying FLUSH
> 0xea Emask 0x12
> Jan 19 21:10:50 enigma kernel: [ 3074.671374] ata2: EH complete
>
> Background info: Gentoo Linux x86_64. I have two almost brand-new 500GB
> drives running in a RAID1 (one of those is a replacement disk for that
> Seagate drive that went bad a few weeks ago, as I mentioned previously),
> with plenty of space on all file systems. There is plenty of unused RAM
> memory while all this is happening, and the CPUs never seem to be
> overloaded.
>
> (For those who are bound to ask, I /stopped/ doing my disk-to-RAM
> experiments several days ago, so that shouldn't be related. :-)
>
> It looks from the log like I might be losing a SATA controller. Any
> thoughts?
>
>
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Wed Jan 19 21:46:03 2011

This archive was generated by hypermail 2.1.8 : Wed Jan 19 2011 - 21:46:03 AKST