Board index HPUX

Help: SCSI NCR stalls on HP 9000/735

Help: SCSI NCR stalls on HP 9000/735

Post by Seth M. LaFor » Fri, 02 Feb 1996 04:00:00



We've been having problems with SCSI stalls recently on our main
server, an HP 9000/735 running HPUX 9.01.  Every 5-30 minutes all SCSI
activity on the machine will halt for about 15 seconds.  syslog logs
the following kern.info messages:

  Jan 26 21:35:02 envy syslog:
  Jan 26 21:35:02 envy syslog: Jan 26 21:35
  Jan 26 21:35:02 envy syslog: SCSI: NCR stall (normal condition), bus = 0x1
  Jan 26 21:55:01 envy syslog:
  Jan 26 21:55:01 envy syslog: Jan 26 21:55
  Jan 26 21:55:01 envy syslog: SCSI: NCR stall (normal condition), bus = 0x1
  Jan 26 21:55:01 envy last message repeated 2 times

As far as I can remember, this behavior first showed up when we added
a new disk, a Quantum XP34300, to the bus.  We've since added another
XP34300.  It seems to me that the timeouts are getting more frequent
as time goes on, but it's hard to tell.

Does anyone have any idea what these errors are, and how to fix them?
Does anyone know of incompatibilities between the Quantum XP34300 and
HP's SCSI hardware?  The XP34300's are in APS cases with Digital
Active Termination - the drive on the end of the bus is an XP34300
with its termination turned on.  Perhaps DATerm is a bad idea?

I searched the HP patch database, and came up with patch PHKL_6050
(s700 9.0[357] SCSI, NFS, VM, HIL, graphics kernel megapatch), which
makes claims to solve this kind of thing:
   PHKL_4099:
      This patch fixes two known corner cases on the NCR53C700
      Single-Ended bus that create timing windows in the scsi_c700
      interface driver. ... The fix was for the driver to check for
      both these conditions and as a result, it addresses all known
      SCSI request timeout problems encountered to date.
Unfortunately, this patch is not for 9.01.  PHKL_6049 claims to be the
equivalent to PHKL_6050 for 9.01 ("Equivalent Patches: PHKL_6050:
s700: 9.03 9.05 9.07"), but doesn't mention SCSI driver fixes.
Indeed, I tried installing it and it hasn't fixed the problems.

Does anyone know if my problem is in fact the one that PHKL_4099
claims to fix?

Here's pertinent system info:

  # uname -a
  HP-UX envy A.09.01 A 9000/735 2010594854 two-user license
  # cd /dev/rdsk ; for drive in *; do echo ; /etc/diskinfo $drive ; done

  SCSI describe of c201d0s0:
               vendor: TOSHIBA
           product id: CD-ROM XM-3301TA
                 type: CD-ROM
                 size: 208304 Kbytes
     bytes per sector: 2048

  diskinfo: can't open c201d1s0: Invalid argument

  diskinfo: can't open c201d2s0: No such device or address

  SCSI describe of c201d3s0:
               vendor: HP      
           product id: 97560          
                 type: direct access
                 size: 1323540 Kbytes
     bytes per sector: 512

  SCSI describe of c201d4s0:
               vendor: Quantum
           product id: XP34300        
                 type: direct access
                 size: 4199760 Kbytes
     bytes per sector: 512

  SCSI describe of c201d5s0:
               vendor: Quantum
           product id: XP34300        
                 type: direct access
                 size: 4199760 Kbytes
     bytes per sector: 512

  SCSI describe of c201d6s0:
               vendor: SEAGATE
           product id: ST3600N        
                 type: direct access
                 size: 511742 Kbytes
     bytes per sector: 512
  # /etc/ioscan -f
  Class      H/W Path  Driver   H/W Status    S/W Status
  =======================================================
  graphics   1.0.0     graph3   ok(0x577)     ok        
  scsi       2.0.1     c700     ok(0x7071)    ok        
  disk       2.0.1.0.0 scsi     ok(0x5800202) ok        
  tape_drive 2.0.1.1.0 scsitape ok(0x1800202) ok        
  disk       2.0.1.3.0 scsi     ok(0x201)     ok        
  disk       2.0.1.4.0 scsi     ok(0x202)     ok        
  disk       2.0.1.5.0 scsi     ok(0x202)     ok        
  disk       2.0.1.6.0 scsi     ok(0x202)     ok        
  lan        2.0.2     lan01    ok(0x7072)    ok        
  hil        2.0.3     hil      ok(0x7073)    ok        
  serial     2.0.4     asio0    ok(0x7075)    ok        
  serial     2.0.5     asio0    ok(0x7075)    ok        
  parallel   2.0.6     parallel ok(0x7074)    ok        
  scsi       2.0.7     c700     ok(0x707c)    ok        
  audio      2.0.8     audio    ok(0x7f)      ok        
  lan        4.1.0     lan01    ok(0x7650)    ok        

Thanks for any help y'all can provide!

Seth

 
 
 
Top

Help: SCSI NCR stalls on HP 9000/735

Post by Doug Siebe » Sat, 03 Feb 1996 04:00:00



Quote:>We've been having problems with SCSI stalls recently on our main
>server, an HP 9000/735 running HPUX 9.01.  Every 5-30 minutes all SCSI
>activity on the machine will halt for about 15 seconds.  syslog logs
>the following kern.info messages:
>  Jan 26 21:35:02 envy syslog:
>  Jan 26 21:35:02 envy syslog: Jan 26 21:35
>  Jan 26 21:35:02 envy syslog: SCSI: NCR stall (normal condition), bus = 0x1
>  Jan 26 21:55:01 envy syslog:
>  Jan 26 21:55:01 envy syslog: Jan 26 21:55
>  Jan 26 21:55:01 envy syslog: SCSI: NCR stall (normal condition), bus = 0x1
>  Jan 26 21:55:01 envy last message repeated 2 times
>As far as I can remember, this behavior first showed up when we added
>a new disk, a Quantum XP34300, to the bus.  We've since added another
>XP34300.  It seems to me that the timeouts are getting more frequent
>as time goes on, but it's hard to tell.

I had similar problems on a 9.01 system, and called it in to HP.  I was told
that my Maxtor MXT1240S drive was unsupported, and nothing could be done
about it.  OK, fine, I just lived with it.  I recently installed 10.01 on
that machine, and put it on a 2G drive that *is* supported (it came in an HP
715/80, but I can't imagine it would be supported in a 715/80 but not in a
710)  With all current patches on 10.01, I *still* have the NCR stalls.  I
suppose I'll call it in at some point, but I've got a bunch of outstanding
calls against 10.01 for more important things right now so I'm going to wait
a bit on it.  The Maxtor is still inside the machine (though it isn't even
mounted at the moment) so I suppose its possible that its presence on the
bus is causing this, I will probably remove it when I get a chance and if I
still see these problems, I will most definitely be opening another call to
HP about this, because it has me very annoyed that I was told it was from
using an unsupported disk and I'm still seeing it with the latest OS, all
patches, and a supported disk.

--
Doug Siebert              || "Usenet is essentially Letters to the Editor
University of Iowa        ||  without the editor.  Editors don't appreciate

(c) 1996 Doug Siebert.  Redistribution via the Microsoft Network is prohibited.

 
 
 
Top

1. HELP! SCSI: NCR stall (normal condition), bus = 0x11

Colleagues,

I have an HP 9000/750 with HP-UX 9.01 and PHKL_4053, with 2 EISA
FAST SCSI-2 cards driving striped Seagate ST41651NDs which perform
well:
     --Sequential Output-- ---Sequential Input-- --Random--
     -Per Char- --Block--- -Per Char- --Block--- --Seeks---
  MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
 100  1459 62.8  1501  8.3  1887 93.5  5607 48.2  58.7  5.0

and striped Seagate ST12400NDs, whis sometimes perform even better:

     --Sequential Output-- ---Sequential Input-- --Random--
     -Per Char- --Block--- -Per Char- --Block--- --Seeks---
  MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
 100  2111 98.8  6408 83.9  1889 97.4  7452 92.6  85.0  7.7

until the SCSI driver registers an NCR stall error via dmesg:

        SCSI: NCR stall (normal condition), bus=0x11

and the disc subsystem (including swap) hangs for 20 seconds causing
all sorts of NFS and rcmd timeouts, not to mention much frustration.

Would anyone know what this error means, and how I might go about
correcting it?

Thanks ever so much in advance,

Sean Batt
--

-------- Coombs Computing Section - Telephone: +61 6 249 3296 -----
-- Australian National University - GPO Box 4 Canberra City 2601 --
-------------------------------------------------------------------

2. sending a message from a Web service to a Windows client

3. SCSI: NCR stall (normal condition), bus = 0x1

4. aiutooooo sono disperato

5. PRISON from actionware

6. Anyone have SCSI floppy in HP 9000 735?

7. PC Games Offer

8. 9000/7x0 UX9.03 NCR SCSI problems

9. FS: HP 9000 735/125 ; 735/100 ; 715/100XC

10. NCR stall -- normal condition?

11. What is "NCR Stall"?

12. NCR stall on 715/33


All times are UTC
Board index