RAID 4Si Contoller Firmware Upgrade

[WEILAND]

Bonjour,
I want to change the revision of the firmware U.01.01 to U.01.06
On an A500 Server running HPUX 11.00, to the card are connected two disc bays DS2100.
I dismounted the files systems and stopped the applications that turned on these file systems.

And with each test I have this error:
# /usr/sbin/i2outil /dev/iop0 download PRIV a5856a_u_01_06.img
Download starting - do NOT cycle power

error in SW_DOWNLOAD ioctl: I/O error
------------------------------------------
There are two possible causes of this error.

Cause #1
Applications are running which have opened the device files pertaining
to the controller or filesystems are mounted on its logical drives.

Action #1
Terminate applications using the controller, unmount any filesystems
using the logical drives, and try the download again.

Cause #2
A download has already been completed, and the driver is in
SW_DOWNLOAD state. The get_stats option of i2outil
will report IOP Driver State of SW_DOWNLOAD in this case.

Action #2
The driver state can only be changed to READY by rebooting the system.
------------------------------------------
#

I think I have omitted something but what?
I have follow stats with i2outil:
# i2outil /dev/iop0 get_stats
****************** IOP STATISTICS ********************
------------------------------------------------------
STATE INFO
==========
IOP Driver State = READY/IN_USE
------------------------------------------------------
MESSAGE SIZE
============
IOP Request Message Size = 128
IOP Reply Message Size = 128
------------------------------------------------------
MESSAGES SENT AND RECEIVED
==========================
Pull Messages Sent = 0, Push Messages Sent = 1354
Total Messages Sent = 1354
Total messages received = 1348
------------------------------------------------------
MISCELLANEOUS INFORMATION
=========================
Number of Spurious Interrupts = 2
Outbound FIFO empty count = 0
#

Thanks for support
Regards
Stéphan

[TwoProc]

You may need to stop measure ware first. I've seen this before in a similar situation - give it a try - can't hurt anything...

Measureware is the daemon that runs around getting stats on your system for glance and perfview - you don't need it all to run. Only to see what your computer is/was doing from a performance standpoint (in general).

/sbin/mwa stop
swinstall -s [yourpatch]
/sbin/mwa start

[Ron Lawson_1]

It does sound like something still has the bus open. There is one sure way to get around this -- shut down the system, unplug the SCSI cable, restart, upgrade the firmware, shutdown, re-attach the SCSI cable, and restart. But clearly your uptime will be affected by this procedure.

[WEILAND]

They are a volume group using the disks from Disk Bays.
Solution was to deactivate the volume group.
# vgchange â a n /dev/datas
Shutting down server, restarting server without starting disk arrays may cause a raid error (flag error on each disk.)
Regards
And happy New Year to you.

Thanks to

[WEILAND]

They are a volume group using the disks from Disk Bays.
Solution was to deactivate the volume group.
# vgchange –a n /dev/datas
Shutting down server, restarting server without starting disk arrays may cause a raid error (flag error on each disk.)
Regards
And happy New Year to you.