Hi there,
I have 3 identical systems running RedHat 5.2 with kernel 2.0.36. serving
some web pages. OK, a LOT of webpages... ;-) but the pages themselves are
served over NFS, and not on the machines themselves.
The system use Symbios SCSI controller (integrated) which is using
ncr53c875 and One 9GB Quantum Atlas III Ultra-2 SCSI disk (operating as
ultrawide).
Lately I've been getting messages of the following sort from syslog:
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: QUEUE FULL! 1 busy, 6
disconne
cted CCBs
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: tagged command queue
depth set
to 6
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: QUEUE FULL! 1 busy, 5
disconne
cted CCBs
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: tagged command queue
depth set
to 5
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: QUEUE FULL! 1 busy, 4
disconne
cted CCBs
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: tagged command queue
depth set
to 4
Nov 8 04:52:35 bronto kernel: ncr53c875-0-<0,0>: QUEUE FULL! 1 busy, 3
disconne
cted CCBs
More accurately - On one system, I see no such messages, on the other I
see them sometimes and on the third I see them almost constantly (but in
bursts), up to a point where the system freezes (The load on these 3
systems is the identical btw).
The driver default max tagged command queue has been changed from the
default of 8 to 15 (as is suggested for large HD/large loads).
According to the driver docs, the QUEUE FULL message, is RARELY emitted
from some SCSI controllers. The driver then lowers the max tagged command
queue (by shortening the queue by the number of "lost" commands) and then
slowly raises it up again every 1000 commands until it reaches the max
tagged command queue. Well, in my case this does NOT happen rarely at all
and sometimes freezes the system.
Anyone knows anything about this? is this a bad HD? a bad SCSI
controllers? hostile alien activity?
Some more info on the machines are at the end of this message. Any extra
information needed - just ask... ;-)
Thanks!
Gilad.
Red Hat linux 5.2 fully installed and configured.
One Intel Pentium II 350MHz processor.
512K ECC L2 cache.
Intel N440BX motherboard.
128MB 100MHz ECC SDRAM.
One 9GB Quantum Atlas III Ultra-2 SCSI disk (operating as ultrawide).
Symbios SCSI controller (integrated).
And the kernel says:
Nov 9 08:31:24 bronto kernel: ide: i82371 PIIX (Triton) on PCI bus 0
function 1
45
Nov 9 08:31:24 bronto kernel: ide0: BM-DMA at 0xfcd0-0xfcd7
Nov 9 08:31:24 bronto kernel: ide1: BM-DMA at 0xfcd8-0xfcdf
Nov 9 08:31:24 bronto kernel: Floppy drive(s): fd0 is 1.44M
Nov 9 08:31:24 bronto kernel: FDC 0 is a National Semiconductor PC87306
Nov 9 08:31:24 bronto kernel: md driver 0.36.3 MAX_MD_DEV=4, MAX_REAL=8
Nov 9 08:31:24 bronto kernel: ncr53c8xx: at PCI bus 0, device 13,
function 0
Nov 9 08:31:24 bronto kernel: ncr53c8xx: 53c875 detected
Nov 9 08:31:24 bronto kernel: ncr53c8xx: at PCI bus 0, device 13,
function 1
Nov 9 08:31:24 bronto kernel: ncr53c8xx: 53c875 detected
Nov 9 08:31:24 bronto kernel: ncr53c875-0: rev=0x37, base=0xfedfec00,
io_port=0
xf800, irq=10
Nov 9 08:31:24 bronto kernel: ncr53c875-0: ID 7, Fast-20, Parity
Checking
Nov 9 08:31:24 bronto kernel: ncr53c875-0: on-chip RAM at 0xfedff000
Nov 9 08:31:24 bronto kernel: ncr53c875-0: restart (scsi reset).
Nov 9 08:31:24 bronto kernel: ncr53c875-0: Downloading SCSI SCRIPTS.
Nov 9 08:31:24 bronto kernel: ncr53c875-1: rev=0x37, base=0xfedfe800,
io_port=0
xf400, irq=10
Nov 9 08:31:24 bronto kernel: ncr53c875-1: ID 7, Fast-20, Parity
Checking
Nov 9 08:31:24 bronto kernel: ncr53c875-1: on-chip RAM at 0xfedfd000
Nov 9 08:31:24 bronto kernel: ncr53c875-1: restart (scsi reset).
Nov 9 08:31:24 bronto kernel: ncr53c875-1: Downloading SCSI SCRIPTS.
Nov 9 08:31:24 bronto kernel: scsi0 : ncr53c8xx - revision 3.0i
Nov 9 08:31:24 bronto kernel: scsi1 : ncr53c8xx - revision 3.0i
Nov 9 08:31:24 bronto kernel: scsi : 2 hosts.
Nov 9 08:31:24 bronto kernel: Vendor: QUANTUM Model: QM39100TD-SW
Rev:
N1B0
Nov 9 08:31:24 bronto kernel: Type: Direct-Access
ANSI
SCSI revision: 02
Nov 9 08:31:24 bronto kernel: Detected scsi disk sda at scsi0, channel
0, id 0,
lun 0
Nov 9 08:31:24 bronto kernel: ncr53c875-0-<0,0>: tagged command queue
depth set
to 15
Nov 9 08:31:24 bronto kernel: scsi : detected 1 SCSI disk total.
Nov 9 08:31:24 bronto kernel: ncr53c875-0-<0,*>: FAST-20 WIDE SCSI 40.0
MB/s (5
0 ns, offset 16)
Nov 9 08:31:24 bronto kernel: SCSI device sda: hdwr sector= 512 bytes.
Sectors=
17783249 [8683 MB] [8.7 GB]
=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
