AFAIK the values must be exact multiply of and not just over the values
stated. Especially the max_coalesce parameter might cause some grief.
Their description is in Appendix D of redbook "Monitoring and Managing IBM
SSA Disk Subsystems" (SG24-5251-00). The book contains also other useful
hints.

To lower them you have two options:
-       varyoffvg against the group using those PV hdisks
-       if not possible (usually for rootvg) - use chdev with -P flag (or
corresponding smitty option) and reboot. The flag instructs AIX to behave
like Windows - update the value in the ODM and wait for reboot to use it.

Zlatko Krastev
IT Consultant






Zoltan Forray/AC/VCU <zforray AT VCU DOT EDU>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
16.01.2003 16:08
Please respond to "ADSM: Dist Stor Manager"


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        Re: Problems with TSM AIX server V5.1.5.4


These values are actually higher than you suggested.

1.  queue depth is set to 48 versus computed value of 45.
2.  max_coalesce is set to x10000 versus computed xF0000

Does it matter that the values are higher ?

Also, when we tried to lower them, we did get the "drive is busy" message
so I am open to your suggestions !!!






Dave Canan <ddcanan AT ATTGLOBAL DOT NET>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
01/15/2003 08:41 PM
Please respond to "ADSM: Dist Stor Manager"


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        Re: Problems with TSM AIX server V5.1.5.4


Can you do the following and re-post with what you have set for the
following?

smitty devices -> SSA disks -> SSA Logical Disks -> Change/Show
Characteristics of an SSA Logical Disk -> select hdiskxx (where xx is the
hdisk where you are defining the volume)

Look at the following fields:

         1. queue depth.      This should be set to (# of disks in
array-1)*3

         2. max_coalesce   This should be set to (# of disks in array
-1)*65536, converted to hex.

You may have trouble setting these devices  - they may report "device is
busy" if you try and change them. If so, re-post and I'll suggest ways to
get around this. We had one customer report numerous errors like this and
after changing these values, the errors went away.


At 03:58 PM 1/15/2003 -0500, you wrote:
>Just tried formatting a new storage pool volume on the above system.
First
>time since upgrading to 5.1.5.x.
>
>Seeing lots of errors:  LVDD LVM_IO_FAIL is the message.
>
>Nothing in ERRPT but these errors. Nothing to point to a specific
hardware
>failure. Physically checked the drives (SSA) and can see any physical
>indication of problems.
>
>Tried searching IBMLINK with no success.
>
>
>Anyone else seen this ?

Dave Canan
TSM Performance
IBM Advanced Technical Support
ddcanan AT us.ibm DOT com