Is there a supported way to attach a Tape Drive to OpenVMS 9.2-2

[pocketprobe]

Are there presently any supported ways to attach a(n LTO) tape drive to a system running 9.2-2?
Ideally via SAS/SCSI

[mberryman]

I am running ESXi 7.0.3n with HPE addons.

ESXi reports the firmware version as 9.03.40
VMS reports the firmware version as 8.08.05

In checking, 8.08.05 is the primary flash and 9.03.40 is the secondary flash. Primary is currently active.

Does VMS have the ability to tell the HBA to switch which flash is active?

[m_detommaso]

If I recall correctly, the firmware running on an HBA is part of the VMS system driver data structure; the operating system does not use the firmware loaded into the adapter's NVRAM.

The below data was captured on a c7000 + Integrity Server BL860ci2 + QLogic QMH2562 8Gb FC HBA :

Code: Select all

2,0,0,0 3404008101E10000 000000000007000C EVN_IO_DISCOVERY_START
Dual Port Flex10 10GbE BL8XXc i2 Embedded CNIC is detected
Dual Port Flex10 10GbE BL8XXc i2 Embedded CNIC is detected
Dual Port Flex10 10GbE BL8XXc i2 Embedded CNIC is detected
Dual Port Flex10 10GbE BL8XXc i2 Embedded CNIC is detected
HP PCIe 2Port 8Gb Fibre Channel Adapter (driver 2.27, firmware 5.06.006)    <-
HP PCIe 2Port 8Gb Fibre Channel Adapter (driver 2.27, firmware 5.06.006)    <-
HP Smart Array P410i Controller     (version 6.64)  1 Logical Drive
Currently the controller is in RAID mode
2,0,0,0 5404020B01E10000 0000000000000006 EVN_EFI_LAUNCH_BOOT_MANAGER

Code: Select all

 Overview
 --------
 Context ........ Live from xxxxxx (10.110), 26-NOV-2024 11:33:17.67
 Device ......... FGA0:, PCI, Device ID 2532, Rev 02, QLogic ISP2532
 Images ......... FW 4.04.04, SYS$PGQDRIVER X-55A1 linked 26-MAY-2024 07:07:15.74   <-
 FibreChannel ... Node WWID 5001.4380.120C.F9AD, Port WWID 5001.4380.120C.F9AC, Port ID 030100

Code: Select all

Overview
 --------
 Context ........ Live from xxxxxx (10.110), 26-NOV-2024 12:11:16.58
 Device ......... FGB0:, PCI, Device ID 2532, Rev 02, QLogic ISP2532
 Images ......... FW 4.04.04, SYS$PGQDRIVER X-55A1 linked 26-MAY-2024 07:07:15.74   <-
 FibreChannel ... Node WWID 5001.4380.120C.F9AF, Port WWID 5001.4380.120C.F9AE, Port ID 040100

I ask for confirmation at VSI engineering anyway

/Maurizio

[roberbrooks]

> If I recall correctly, the firmware running on an HBA is part of the VMS system driver data structure; the
> operating system does not use the firmware loaded into the adapter's NVRAM.

That is no longer the case.

That had been the behaviour, but that changed for 16Gb/sec adapter support (on both IA64 and X86_64).

Added in 3 minutes 7 seconds:
> ESXi reports the firmware version as 9.03.40
> VMS reports the firmware version as 8.08.05

> In checking, 8.08.05 is the primary flash and 9.03.40 is the secondary flash. Primary is currently active.

> Does VMS have the ability to tell the HBA to switch which flash is active?

That's strange that the primary firmware would appear to be older than the secondary (backup) firmware.

I'm not working this week; I'll check this out with the driver writer (if Paul Jacobi doesn't see this post here first).

[mberryman]

I found an error in the zone configuration of the fiber switches. After correcting that, the tape drives are now visible and working fine. In summary, I have now tested both disk and tape access using the following HBAs:

2692
2660
2532

The only problem I have encountered is that putting the 2660 in passthru mode causes all of the fans in the DL380 to go to 100% speed and stay there. No other card causes this and I haven't been able to figure out why this one does.

[roberbrooks]

> The only problem I have encountered is that putting the 2660 in passthru mode causes all of the fans in the
> DL380 to go to 100% speed and stay there. No other card causes this and I haven't been able to figure out
> why this one does.

We've never seen that, but I'm sure we've never tried passthrough on the 2660.

Whenever we encounter something weird, I always use the latest HPE Proliant firmware update ISO.

I can't remember which card it was, but passthrough would not work with a firmware rev that was only one back from current. Once we updated the firmware, all was well.