ATA problems with 5.5-PRERELEASE

[Alfred Perlstein]

Hello, we recently began deploying 5.5-PRERELEASE dated
March 7th.

Across all (7 so far) of our machines we're getting the following error:

ad0: TIMEOUT - WRITE_DMA retrying (2 retries left) LBA=32176383
ad0: TIMEOUT - WRITE_DMA retrying (2 retries left) LBA=15514623
ad0: TIMEOUT - WRITE_DMA retrying (2 retries left) LBA=34480383
ad0: TIMEOUT - WRITE_DMA retrying (2 retries left) LBA=31408319
ad0: TIMEOUT - WRITE_DMA retrying (2 retries left) LBA=15718783

On FreeBSD 5.4-stable from September we were fine.

This is the ata hardware present:

atapci0: <Intel ICH3 UDMA100 controller> port 0xffa0-0xffaf,0x376,0x170-0x177,0x
3f6,0x1f0-0x1f7 at device 31.1 on pci0
ata0: channel #0 on atapci0
ata1: channel #1 on atapci0
ad0: 38166MB <WDC WD400JB-00ENA0/05.03E05> [77545/16/63] at ata0-master UDMA100
acd0: CDROM <SR244W/T01A> at ata1-master PIO4

>From mysql on one of the hosts:

060309  5:28:48 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'
060309  5:29:40 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'
060309  5:30:02 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'
060309  5:30:14 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'
060309  5:30:14 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'
060309  5:34:40 [ERROR] Got error 134 when reading table './romatch/profile_acti
ve'

#define HA_ERR_RECORD_DELETED   134     /* Intern error-code */

Looks like we were getting corrupt data.

Any hints?

Can this be looked into please?

-- 
- Alfred Perlstein
- CTO Okcupid.com / FreeBSD Hacker / All that jazz -