We have immediate_reporting set to 1 on devices /dev/rdsk/c24t0d0 and /dev/rdsk/c26t0d0 and want to change immediate_reporting to 0 on these devices. We get the following error:
The immediate_reporting value remains 1. It was set to 0 by default when the devices were created. We tried rebooting the machine. How can we force immediate_reporting to 0?
Re: scsictl error when setting immediate_reporting
I only have an HP-UX 11.11 system to look at, and you don't specify which version you are attempting this on, nor what storage system you're addressing, which may also play in this regard.
It doesn't appear that I can change that parameter on my system for SAN disks, though I can for direct-attached disks.
Re: scsictl error when setting immediate_reporting
> ...is an LSI disk array.
Wow, that sure doesn't help much. LSI Corp. has dozens of products from ICs used in drives to HBAs and RAID controllers, etc. scsictl is used only with real disks, not LUNs presented as virtual disks by a smart array controller. Is your LSI array just a bunch of disks in a small cabinet with a simple SCSI cable? Or can you configure the array to specify the size of the disk(s) presented to your system? I suspect that this question would be answered more accurately from LSI.
Re: scsictl error when setting immediate_reporting
RESOLUTION:
We resolved the problem by going into the SANTricity software that is used to manage the array, right-clicking on the LUN, going into the cache settings, and disabling the "write-cache enable" setting. Then, we were successful in setting the immediate_report value to 0 through the scsictl command. We have an LSI SAN connected by two fibre channel cables to a switch and the switch is connected by two fibre channel cables each to our HP-UX servers, which run HP-UX 11.11.
