public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Command-line tools for stress-testing SATA hard disks (and libata)
@ 2008-05-12 22:10 Bradley Chapman
  2008-05-12 22:34 ` Dieter Ries
  0 siblings, 1 reply; 2+ messages in thread
From: Bradley Chapman @ 2008-05-12 22:10 UTC (permalink / raw)
  To: linux-kernel

(Please CC me on replies.)

I have a new(ish) 160GB Maxtor SATA hard disk that recently cracked up 
and failed with lots of ext3 filesystem errors; beforehand, it had 
started to become flaky and was stalling under certain I/O conditions, 
emitting DRDY ABRTs to the libata layer and forcing constant link 
resets. At the moment, the disk appears to be unbootable - GRUB fails to 
load the stage 2 loader and hangs after loading stage1.5.
Unfortunately, these problems were reproduced on tainted 2.6.24 and 
2.6.25 kernels (specifically, fglrx was doing the tainting). Ergo, I'm 
not formally reporting this as a libata problem (yet).

Can anyone here recommend a command-line disk benchmark or stress test 
utility that doesn't need to be run in X (allowing me to test with 
untainted kernels) and would be capable of producing a variety of I/O 
access patterns that could trigger the errors I've seen during normal 
usage? I've since replaced the SATA cable and switched the hard disk to 
a different SATA PHY port, and when I mount individual partitions and 
access files, the libata layer appears to behave.

Brad

P.S: The 2.6.25.3 libata reports the potentially broken disk as follows:

sata_nv 0000:00:0d.0: version 3.5
...
ata4: SATA max UDMA/133 cmd 0x970 ctl 0xb70 bmdma 0xe008 irq 23
...
ata4: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata4.00: ATA-7: MAXTOR STM3160815AS, 3.AAD, max UDMA/133
ata4.00: 312581808 sectors, multi 1: LBA48 NCQ (depth 0/32)
ata4.00: configured for UDMA/133
scsi 3:0:0:0: Direct-Access     ATA      MAXTOR STM316081 3.AA PQ: 0 ANSI: 5
sd 3:0:0:0: [sdb] 312581808 512-byte hardware sectors (160042 MB)
sd 3:0:0:0: [sdb] Write Protect is off
sd 3:0:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 3:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't 
support DPO or FUA
sd 3:0:0:0: [sdb] 312581808 512-byte hardware sectors (160042 MB)
sd 3:0:0:0: [sdb] Write Protect is off
sd 3:0:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 3:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't 
support DPO or FUA
  sdb: sdb1 sdb2 sdb3 sdb4 < sdb5 sdb6 sdb7 sdb8 sdb9 sdb10 >
sd 3:0:0:0: [sdb] Attached SCSI disk

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Command-line tools for stress-testing SATA hard disks (and libata)
  2008-05-12 22:10 Command-line tools for stress-testing SATA hard disks (and libata) Bradley Chapman
@ 2008-05-12 22:34 ` Dieter Ries
  0 siblings, 0 replies; 2+ messages in thread
From: Dieter Ries @ 2008-05-12 22:34 UTC (permalink / raw)
  To: kakadu84; +Cc: linux-kernel

Hi,

Bradley Chapman schrieb:
> Can anyone here recommend a command-line disk benchmark or stress test 
> utility that doesn't need to be run in X (allowing me to test with 
> untainted kernels) and would be capable of producing a variety of I/O 
> access patterns that could trigger the errors I've seen during normal 
> usage? 

Why don't you just use the vesa driver and use your stress test utility 
in X?

> Brad

cu
	Dieter

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2008-05-12 22:50 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-05-12 22:10 Command-line tools for stress-testing SATA hard disks (and libata) Bradley Chapman
2008-05-12 22:34 ` Dieter Ries

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox