The ODM is your last resort. I agree. I would try taking the definitions
out of all of the classes of the ODM for those disks and then run cfgmgr
again or reboot even, to see what happens.
--Justin Richard Bleistein
Unix Systems Administrator (Sungard eSourcing)
Desk: (856) 866 - 4017
Cell: (856) 912 - 0861
Email: [EMAIL PROTECTED]
Todd Lundstedt
<Todd_Lundstedt@VIA-C To: [EMAIL PROTECTED]
HRISTI.ORG> cc:
Sent by: "ADSM: Dist Subject: Re: AIX question
Stor Manager"
<[EMAIL PROTECTED]
>
01/07/2003 10:35 AM
Please respond to
"ADSM: Dist Stor
Manager"
Becky (and Justin from another post),
Both of you suggest that I first delete the offending disks... yet that is
my exact problem. I can't delete them. When I try to "rmdev" them, I get
an error indicating that there is not enough information to uniquely
identify the device (rmdev -l 'hdisk2' -d ... but there are two "hdisk2"
devices). I was looking for a way to delete them other than rmdev.
The AIX guru in our shop finally came up with something he remembered from
his classes... odmget and odmdelete commands. They appear to allow one to
"more uniquely identify" a single object, more so than rmdev. Are there
any gotchas to consider while using odmget and odmdelete?
Todd
"Davidson,
Becky" To: [EMAIL PROTECTED]
<Becky.Davidson cc:
@SLBG.COM> Fax to:
Sent by: "ADSM: Subject: Re: AIX question
Dist Stor
Manager"
<[EMAIL PROTECTED]
ST.EDU>
01/06/2003
06:47 PM
Please respond
to "ADSM: Dist
Stor Manager"
Delete the two hdisk again. Run cfgmgr again. Check it and if you see the
same type of pattern of 1 vpath showing 4 and the other 2 showing only one
then the shark guy needs to recheck his configuration. If you see a
different pattern try running cfgmgr again. There was a problem awhile
back
that you need to run cfgmgr x's with x being the number of paths so that
you
would get all of the paths. Sometimes we have had to delete and run cfgmgr
again because it doesn't seem to work right the first time. After delete
and rerunning cfgmgr it seems to work better. Something you might want to
get your hands on is the Shark Gotcha's document. I think it is still on
the storage page but if not email me off list and I will send it to you.
Another issue you might run into is that the devices may not be fully
formatted. There was a version of the shark code that said that everything
was formatted yet it wasn't. The only way you could tell was to go into
the
shark under where you configure paths and refresh the screen.
good luck and let me know if I can be of further assistance
Becky
-----Original Message-----
From: Todd Lundstedt [mailto:[EMAIL PROTECTED]]
Sent: Monday, January 06, 2003 4:50 PM
To: [EMAIL PROTECTED]
Subject: AIX question
First, let me say..
Dear Lord, the IBM web pages are a pain to dig through and find what you
are looking for. Couldn't find any knowledge bases or anything to help me
resolve the issue I have... so.. here goes.
AIX 4.3.3
My TSM server is a pSeries with two Fibre Channel Adapters installed,
through which we access data on the ESS (shark) and the LTO tape drives (I
know, bad idea). Each of the fibre cards is connected to its own fibre
switch. Each switch has two connections to the Shark (four data paths to
the Shark).
We have been trying to add some additional Shark storage to the TSM server.
After the Shark guy did his thing, I ran cfgmgr, and thinks didn't look
quite right. The two old "vpath" devices each showed four disks (hdsikX),
two on each of the two fibre cards (fscsi1, fscsi0), just like they always
have. One of the three new "vpath" devices showed four hdisks on two
different "fscsi" devices, but the other two "vpath" devices only showed
two hdisks, one on each fscsi.
lscfg showed a bunch of new disks, mostly "available", but four were
"defined". I decided to delete the three new "vpath" devices, and all of
the associated "hdisk" devices, including the defined ones. As I got to
looking, I saw something odd...
++++++++++++++++
root@ax-dco-tsm / # lsdev -Cc disk
hdisk0 Available 10-88-00-8,0 16 Bit LVD SCSI Disk Drive
hdisk1 Available 10-88-00-9,0 16 Bit LVD SCSI Disk Drive
hdisk18 Available 10-68-01 IBM FC 2105F20
hdisk19 Available 10-68-01 IBM FC 2105F20
hdisk20 Available 10-68-01 IBM FC 2105F20
hdisk21 Available 10-68-01 IBM FC 2105F20
hdisk22 Available 10-78-01 IBM FC 2105F20
hdisk23 Available 10-78-01 IBM FC 2105F20
hdisk24 Available 10-78-01 IBM FC 2105F20
hdisk25 Available 10-78-01 IBM FC 2105F20
vpath3 Available Data Path Optimizer Pseudo Device Driver
vpath4 Available Data Path Optimizer Pseudo Device Driver
hdisk2 Defined 10-78-01 IBM FC 2105F20 ****
hdisk2 Defined 10-68-01 IBM FC 2105F20 ****
hdisk3 Defined 10-68-01 IBM FC 2105F20 ****
hdisk3 Defined 10-68-01 IBM FC 2105F20 ****
root@ax-dco-tsm / # datapath query device
Total Devices : 2
DEV#: 0 DEVICE NAME: vpath3 TYPE: 2105F20 SERIAL: 00719516
================================================================
Path# Adapter/Hard Disk State Mode Select
Errors
0 fscsi0/hdisk18 OPEN NORMAL 25
0
1 fscsi0/hdisk20 OPEN NORMAL 31
0
2 fscsi1/hdisk22 OPEN NORMAL 29
0
3 fscsi1/hdisk24 OPEN NORMAL 26
0
DEV#: 1 DEVICE NAME: vpath4 TYPE: 2105F20 SERIAL: 60219516
================================================================
Path# Adapter/Hard Disk State Mode Select
Errors
0 fscsi0/hdisk19 OPEN NORMAL 13465
0
1 fscsi0/hdisk21 OPEN NORMAL 13636
0
2 fscsi1/hdisk23 OPEN NORMAL 13567
0
3 fscsi1/hdisk25 OPEN NORMAL 13727
0
++++++++++++++++++++++
Note the **** lines.
I have two entries for "hdisk2" and two for "hdisk3". Smitty won't let me
remove them (not enough information to uniquely identify the device) and
similar error messages for "change/show charactaristics".
How can I remove "hdisk2" and "hdisk3"? The 10-78 and 10-68 devices are
the fibre cards.
Thanks in advance
Todd Lundstedt
Technical Specialist
Via Christi Information Management Services
ofc. (316) 261-8385
fax. (316) 660-0036
[EMAIL PROTECTED]
