Scsi Reset on syslog.log file ( HPUX11, SC10, MC/SG )

[azzarini]

Configuration :
- A cluster MC/ServiceGuard into HP-UX11 (
( Server Leonardo = HP-L1000
and server Galileo = HP-L3000 )
- Two card Dual port combo Ultra2SCSI (A5838A)
into Galileo
- Two card Dual port combo Ultra2SCSI (A5838A)
into Leonardo
- Two SC10 Disk Array

The first SC10 is connected :
SCSI A vs Leonardo ( 1^ card dual )
SCSI B vs Galileo ( 1^ card dual )
Terminator into SCSI A other port
Terminator into SCSI B other port
The second SC10 is connected :
SCSI A vs Leonardo ( 2^ host-dual card )
SCSI B vs Galileo ( 2^ host-dual card )
Terminator into SCSI A other port
Terminator into SCSI B other port
I use auto-terminated cables and i have setted switch in mode no-termination
into all four host-dual card
I have setted single Initiator via PDC whit different values ( 4,5,6,7 )
When i see the initiator situation :
1. ioscan â fnk | grep Initiator
in this case all Initiar device is 7
ctl 20 0/1/0/0/4/0.7.0 sctl CLAIMED DEVICE Initiator
ctl 23 0/1/0/0/4/1.7.0 sctl CLAIMED DEVICE Initiator
ctl 26 0/9/0/0/4/0.7.0 sctl CLAIMED DEVICE Initiator
ctl 29 0/9/0/0/4/1.7.0 sctl CLAIMED DEVICE Initiator
2. From PDC the 0/9/0/0 channel is correctly
setted to 4
>SER >SCSI
0/0/1/0 7 Ultra Unknown
0/0/1/1 7 Ultra Unknown
0/0/2/0 7 Ultra Unknown
0/0/2/1 7 Ultra Unknown
0/1/0/0 7 No Limit Unknown
0/9/0/0 4 No Limit Unknown


Problem :
I have continuos SCSI Reset in syslog.log file

Questions :
1. Is it necessary to use the terminator into 2^ SCSI host-dual port ?

Thanks for your help.
Marco Azzarini
mail : marco.azzarini@dp2000.

[Steven E. Protter]

SCSI resets may be nothing, they may be a sign of future trouble.

I would be interested in seeing if EMS reports any problems with SCSI.

Break out mstm or xstm(Xwindows) and exercize the devices and see if anything comes up with problems.

Our rp5450 servers do not have terminators on their very similar(same part number I think) port. They are apparently self terminating. When we hung Ultrium drives off our LVD SCSI cards they were self terminating. I generated a service call by terminating it myself(Doh!).

SEP

[RolandH]

1.I reccomend first only attach one system to both SC10.
Do you see the disks in the PDC if you do a SEA?

2. Try again to set the initiator id on your Interfaces.
go into service menu
> scsi init 0/0/1/0 6
> scsi init 0/0/2/0 6

Connect these ports to SC10 one and two.
Be aware that you use now id 6 for the initiator. So you must remove the disk, which is in slot 7 (id 6). Otherwise you have a new id conflict. Do know see the disks from both systems?

3. Connect second system to both SC10. You can leave here the id 7 if you connect the hw path as from first system I mentioned above.

4. Do you see now all the disks?

HTH
Roland

[Bernhard Mueller]

Marco,

as I see it you have not correctly set the SCSI IDs:
you need to specify e.g.
0/9/0/0/4/0 init 5
0/9/0/0/4/1 init 4

Regards,
Bernhard

[azzarini]

Thanks
I have resolved whit a correct setting of initiator !

From PDC -> SERVICE
scsi init 0/9/0/0/4/1 6
scsi init 0/9/0/0/4/0 6
scsi init 0/1/0/0/4/1 7
scsi init 0/1/0/0/4/0 7

After this operation the ioscan command is correct and the syslog.log file is whitout
'scsi reset' errors

MarcoAzzarini