gmirror/HW problem

[Jiri Mikulas]

Ahoj
mam trochu problem s gmirrorem.
mam 1U od Intelu s deskou S3000AH

FreeBSD 6.2-STABLE FreeBSD 6.2-STABLE #1: Thu Aug  9 12:08:52 CEST 2007   /usr/obj/usr/src/sys/SMP  i386

CPU: Intel(R) Core(TM)2 CPU          6320  @ 1.86GHz (1870.48-MHz 686-class CPU)
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
ACPI APIC Table: <INTEL  S3000AH>
atapci1: <Intel ICH7 SATA300 controller> port 0x30c8-0x30cf,0x30e4-0x30e7,0x30c0-0x30c7,0x30e0-0x30e3,0x30a0-0x30af mem 0x88200000-0x882003ff irq 19 
at device 31.2 on pci0

Deje se mi zvlastni vec
kdyz vytvorim gmirror vlozim do nej druhy disk a dam rebuild tak mi po case disk vypadne.

Sep 15 23:35:23 s1n kernel: GEOM_MIRROR: Device gm0: provider ad4 detected.
Sep 15 23:35:23 s1n kernel: GEOM_MIRROR: Device gm0: rebuilding provider ad4.
Sep 16 03:31:56 s1n kernel: ad4: TIMEOUT - WRITE_DMA48 retrying (1 retry left) LBA=379913728
Sep 16 03:32:08 s1n kernel: ad4: TIMEOUT - WRITE_DMA48 retrying (0 retries left) LBA=379913728
Sep 16 03:32:08 s1n kernel: ad4: FAILURE - WRITE_DMA48 timed out LBA=379913728
Sep 16 03:32:08 s1n kernel: GEOM_MIRROR: Synchronization request failed (error=5). ad4[WRITE(offset=194515828736, length=131072)]
Sep 16 03:32:08 s1n kernel: GEOM_MIRROR: Device gm0: provider ad4 disconnected.
Sep 16 03:32:08 s1n kernel: GEOM_MIRROR: Device gm0: rebuilding provider ad4 stopped.

Aug 20 17:11:20 s1n kernel: GEOM_MIRROR: Device gm0: rebuilding provider ad6.
Aug 20 22:22:44 s1n kernel: GEOM_MIRROR: Device gm0: rebuilding provider ad6 finished.
Aug 20 22:22:44 s1n kernel: GEOM_MIRROR: Device gm0: provider ad6 activated.
Aug 21 16:46:04 s1n kernel: ad6: TIMEOUT - WRITE_DMA48 retrying (1 retry left) LBA=457315935
Aug 21 16:46:16 s1n kernel: ad6: TIMEOUT - WRITE_DMA48 retrying (0 retries left) LBA=457315935
Aug 21 16:46:16 s1n kernel: ad6: FAILURE - WRITE_DMA48 timed out LBA=457315935
Aug 21 16:46:16 s1n kernel: GEOM_MIRROR: Request failed (error=5). ad6[WRITE(offset=234145758720, length=131072)]
Aug 21 16:46:16 s1n kernel: GEOM_MIRROR: Device gm0: provider ad6 disconnected.
Sep  4 11:53:38 s1n kernel: ad6: detached


Nekolikat vypadl i ad6, ale ad4 vypadaval casteji.
Dokud jsme mel server doma na stole se Seagatama, tak to i pod simulovanou zatezi vsechno jelo v poradku.
Disky drzely ok, gmirror taky. Po prevezeni do serverovny a zamontovani do racku se to zacalo chovat takhle divne.
Server jeste neni v ostrem provozu, takze zatez je na nem minimalni.

co uz jsem udelal - v uvedenem poradi:
~~~
  vymenil oba disky, nejdriv jsem tam mel
   ad4: 305245MB <Seagate ST3320620AS 3.AAK> at ata2-master SATA150
   ad6: 305245MB <Seagate ST3320620AS 3.AAJ> at ata3-master SATA150
  nyni tam mam
   ad4: 305245MB <WDC WD3200YS-01PGB0 21.00M21> at ata2-master SATA150
   ad6: 305245MB <WDC WD3200YS-01PGB0 21.00M21> at ata3-master SATA150
  u obou typu/znacek se to chova stejne,

  upgrade biosu na posledni
   Version: S3000.86B.02.00.0044.071120071047
   Release Date: 07/11/2007
~~~~~

co jsem NEudelal:
vymena SATA kabelu
zkusit jiny SATA radic/externi PCI kartu
vymena desky - prodejce umoznuje, ale musi mit podklady na reklamaci, ze to je 100% HW zavada

Setkali jste se s tim nekdo?
Mate nejaky napad co jeste zkusit, jaka by mohla byt pricina?
Ja mam podezreni na desku...
Neexistuje nejaky nastroj pro otestovani radice jako takoveho?

Dik za napady
guli