HP 2920-24G (J9726A) Firmware Questions...

[mc1903-1]

Hi,

I have a few questions regarding firmware on the HP2920-24G's (J9726A), that I cannot find references to in the manuals...

1) Does upgrading the firmware in a stacked configuration upgrade the firmware on ALL stacked switches or just the commander..?

2) After a stacked firmware upgrade, do all switches reboot at the sametime or are they rebooted in sequence (i.e commander first) 

3) Can I 'downgrade' firmware in a stacked configuration. I have a stack that is currently running with an Early Availability release (WB.15.14.0002) in order to get the Redundant Power Supply (RPS) support. This stack is experiancing random member reboots with " chassis: ST1-CMDR: Member 2/1-24 Software exception at bsp_interrupts.c:90 -- in 'fault_handler', task ID - 0x1fb01d40 ". Now I see there is a General Availability release (WB.15.13.0005) and I wonder of this is a more stable release..?

Any thoughts, suggestions or comments will be appreciated...

Thanks

Martin

[UG_Gallagher]

I just got the same error.

00732 chassis: ST1-CMDR: Member 3/1-48 Software exception at bsp_interrupts.c:90 -- in 'fault_handler', task ID = 0x1fb01d40 -> Master Induced Crash. IGNORE!

I amd on firmware version WB.15.16.0005.

Is there a fix.  What is a good version to go with?

[Tom Lyczko]

I think you have to install firmware individually on each switch regardless of stacking.

I have always done it this way with our stacked switches, which I think are all configured to 'candidate'...

Did you open a case with HP support to ask them about this and your firmware issue??
Thank you, Tom

[robidan]

Just tested this with a stack of two 2940-24's, updating from WB.15.17.0007 to WB.15.18.0006.

- Uploaded the new image to the stack and issued the reload command via CLI.

- Commander rebooted itself twice and then got stuck in test mode (all LEDs on).

- After waiting for 15 minutes, connected via serial console to Commander to see what's going on, but device seemed dead.

- Manually powercycled Commander and watched console output. Nothing, until switch was up (again in test mode, with all LEDs on), then was greeted with login prompt.

- After login, it became clear that I was connected to Standby switch, which took over and still ran the previous firmware.

- Issued reboot command via CLI, which - obviously - reset the Standby switch and took the complete stack offline (duh).

- While Standby switch booted, Commander also restarted.

- After a couple of minutes, both switches were back online, running the lastest firmware.

Luckily this was on a non-production stack and I was just testing. Not sure if I did the wrong thing (likely), or if the process is flawed. 

[BrianDavis]

After you upload the new firmware to primary in flash you need to use the 'boot system' command.  This will reboot both members and they will be on the new primary flash.

[Johnny_Dempsey]

Was this ever resolved with a firmware upgrade?