Date: Mon, 27 Dec 1999 14:13:04 +0300 From: Alexander Bezroutchko <abb@zenon.net> To: Tom <tom@sdf.com> Cc: scsi@freebsd.org Subject: Re: IFT3102 and FreeBSD 3.4-STABLE troubles Message-ID: <386749C0.34EBADE5@zenon.net> References: <Pine.BSF.4.05.9912241620520.18844-100000@misery.sdf.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Tom wrote:
> Well, I'm testing a single IFT-3102U2G on a dual-PIII under 3.4 stable.
> I'm assuming that the take-over by the redundant controller is similar
> in appearance to the host, as resetting a controller.
Yes, but take-over takes 4 sec, and resetting takes about 1 minute.
> I've done a few resets of the IFT controller until full load (three
> instances of postmark). FreeBSD paused, then printed a bunch of errors
> and then continued. I have had a incident where FreeBSD just hung after
> resetting the controller. I couldn't reproduce it though.
I've done a lot of resets and all of them lead to crash or inoperability ;(.
For example, take a look on following console snapshots:
Example N1:
~~~~~~~~~~~
Host has no local storage, ahc0 connected to IFT (swap resides on da0b):
-- before controller reset ---
box2# dmesg
...
ahc0: <Adaptec aic7896/97 Ultra2 SCSI adapter> rev 0x00 int a irq 19 on
pci0.12.0
ahc0: aic7896/97 Wide Channel A, SCSI Id=7, 16/255 SCBs
ahc1: <Adaptec aic7896/97 Ultra2 SCSI adapter> rev 0x00 int a irq 19 on
pci0.12.1
ahc1: aic7896/97 Wide Channel B, SCSI Id=7, 16/255 SCBs
...
SMP: AP CPU #1 Launched!
da0 at ahc0 bus 0 target 0 lun 0
da0: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da0: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da0: 5120MB (10485760 512 byte sectors: 255H 63S/T 652C)
da1 at ahc0 bus 0 target 0 lun 1
da1: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da1: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da1: 10000MB (20480000 512 byte sectors: 255H 63S/T 1274C)
da2 at ahc0 bus 0 target 0 lun 2
da2: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da2: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da2: 5120MB (10485760 512 byte sectors: 255H 63S/T 652C)
da3 at ahc0 bus 0 target 0 lun 3
da3: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da3: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da3: 10000MB (20480000 512 byte sectors: 255H 63S/T 1274C)
changing root device to da0s1a
...
box2# mount
/dev/da0s1a on / (ufs, local, writes: sync 7 async 62)
/dev/da0s1e on /var (ufs, local, writes: sync 100 async 150)
/dev/da1a on /usr/obj (ufs, local, writes: sync 2 async 0)
procfs on /proc (procfs, local)
--- then, I run 'find' and reset controller ---
box2# find / > /dev/null
(da0:ahc0:0:0:0): Invalidating pack
(da0:ahc0:0:0:0): Invalidating pack
Dec 27 12:48:03 box2 /kernel: (da0:ahc0:0:0:0): Invalidating pack
Dec 27 12:48:03 box2 /kernel: (da0:ahc0:0:0:0): Invalidating pack
Dec 27 12:48:35 box2 /kernel: (da0:ahc0:0:0:0): Invalidating pack
Dec 27 12:48:35 box2 /kernel: (da0:ahc0:0:0:0): Invalidating pack
zsh: device not configured: /var/mail/root
box2#
--- here controller is up, but freebsd is broken
box2# ls
spec_getpages: I/O read failure: (error code=6)
size: 65536, resid: 65536, a_count: 65536, valid: 0x0
nread: 0, reqpage: 0, pindex: 0, pcount: 16
spec_getpages: I/O read failure: (error code=6)
size: 65536, resid: 65536, a_count: 65536, valid: 0x0
nread: 0, reqpage: 0, pindex: 0, pcount: 16
Dec 27 12:51:54 box2 /kernel: spec_getpages: I/O read failure: (error
code=6)
zsh: device not configured: ls
Dec 27 12:51:54 box2 /kernel: spec_getpages: I/O read failure: (error
code=6)
Dec 27 12:51:54 box2 /kernel: size: 65536, resid: 65536, a_count: 65536,
valid: 0x0
Dec 27 12:51:54 box2 /kernel: size: 65536, resid: 65536, a_count: 65536,
valid: 0x0
Dec 27 12:51:54 box2 /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 16
zsh: device not configured: /var/mail/root
Dec 27 12:51:54 box2 /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 16
Dec 27 12:51:54 box2 /kernel: spec_getpages: I/O read failure: (error
code=6)
Dec 27 12:51:54 box2 /kernel: spec_getpages: I/O read failure: (error
code=6)
Dec 27 12:51:54 box2 /kernel: size: 65536, resid: 65536, a_count: 65536,
valid: 0x0
Dec 27 12:51:54 box2 /kernel: size: 65536, resid: 65536, a_count: 65536,
valid: 0x0
Dec 27 12:51:54 box2 /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 16
Dec 27 12:51:54 box2 /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 16
box2#
----------------------------------------------------------------
Example N2:
~~~~~~~~~~~
Host has one local disk (da4) connected to ahc1, ach0 connected to IFT
(swap resides on da4b).
--- before controller reset ---
box1# dmesg
...
ahc0: <Adaptec aic7896/97 Ultra2 SCSI adapter> rev 0x00 int a irq 19 on
pci0.12.0
ahc0: aic7896/97 Wide Channel A, SCSI Id=7, 16/255 SCBs
ahc1: <Adaptec aic7896/97 Ultra2 SCSI adapter> rev 0x00 int a irq 19 on
pci0.12.1
ahc1: aic7896/97 Wide Channel B, SCSI Id=7, 16/255 SCBs
SMP: AP CPU #1 Launched!
...
da0 at ahc0 bus 0 target 0 lun 0
da0: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da0: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da0: 5120MB (10485760 512 byte sectors: 255H 63S/T 652C)
da4 at ahc1 bus 0 target 0 lun 0
da4: <SEAGATE ST39173WC 4218> Fixed Direct Access SCSI-2 device
da4: 10.000MB/s transfers (5.000MHz, offset 15, 16bit), Tagged Queueing
Enabled
da4: 8683MB (17783240 512 byte sectors: 255H 63S/T 1106C)
da1 at ahc0 bus 0 target 0 lun 1
da1: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da1: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da1: 10000MB (20480000 512 byte sectors: 255H 63S/T 1274C)
da2 at ahc0 bus 0 target 0 lun 2
da2: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da2: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da2: 5120MB (10485760 512 byte sectors: 255H 63S/T 652C)
da3 at ahc0 bus 0 target 0 lun 3
da3: <IFT 3102 0223> Fixed Direct Access SCSI-2 device
da3: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing
Enabled
da3: 10000MB (20480000 512 byte sectors: 255H 63S/T 1274C)
changing root device to da4s4a
changing root device to da4a
...
box1# mount
/dev/da4a on / (ufs, local, writes: sync 135 async 294)
procfs on /proc (procfs, local)
/dev/da0a on /mnt (ufs, local, writes: sync 2 async 1)
...
--- then I run 'find' and reset controller ---
box1# find /mnt -ls > /dev/null
(da0:ahc0:0:0:0): Invalidating pack
Dec 27 13:37:21 box1 /kernel: (da0:ahc0:0:0:0): Invalidating pack
Dec 27 13:37:21 box1 /kernel: (da0:ahc0:0:0:0): Invalidating pack
(da0:ahc0:0:0:0): Invalidating pack
(da0:ahc0:0:0:0): Invalidating pack
(da0:ahc0:0:0:0): Invalidating pack
(da0:ahc0:0:0:0): Invalidating pack
Dec 27 13:37:33 box1 last message repeated 4 times
Dec 27 13:37:33 box1 last message repeated 4 times
find: /mnt/usr/home: Device not configured
find: /mnt/usr/obj: Device not configured
find: /mnt/usr/games: Device not configured
find: /mnt/usr/ports: Device not configured
find: sys: Device not configured
find: home: Device not configured
box1#
--- here controller is up, but freebsd is broken
box1# ls -la /mnt
box1# umount /mnt
umount: unmount of /mnt failed: Device not configured
box1#
----------------------------------------------------------------
Unfortunately, I have never seen correct behaviour of FreeBSD after
controller reset occured during any activity on filesystem mounted from IFT.
I have seen about 20 messages related to IFT scsi-to-scsi controllers in
this list. Did anybody investigate behaviour of FreeBSD during controller
take-over and reset ?
SY,
Alexander Bezroutchko
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?386749C0.34EBADE5>