Solaris

well, I consoled onto the master SC today. Im getting tons of errors messages to the console, but I was able to try and fail the SC's over.

interrupt ring buffer overflow java.io.IOException
interrupt ring buffer overflow java.io.IOException
interrupt ring buffer overflow java.io.IOException
interrupt ring buffer overflow java.io.IOException

I did several setfailover -y force, after ensuring the slave and master sc's had the setfailover on/enable, checking with showfailover.
# setfailover -y force
SC FAilover is not active

I did the setfailover -y force, and it just came back stating it was disabled ? dispite showfailover displaying it 2wasnt.

If I do a showlogs I get

frame manager monitoring is disabled connection is down


So, I checked the NIC's, and when removing / plugging in, I was getting console messages about missing link, then found link, so they are all working. Im assuming Im getting the connection down error, as the master console is erroring so much that you cant connect on, well, it takes so long to reply, you time out !

So, master sc cant failover over to slave, slave cant take the `master` role from the master. Ive rebooted the slave.

So, do I just press reset on the master SC ? If I do, will the sc just reboot, and come back online as the master ? or failover to the slave ?? Will the domains crash as I havent shut them down and getting time to do that is going to be hard !!

The version of my ScApp is 5.15.X


Cheers

SBK

Anyone confirm my above questions ?

SBK

I would expect downtime to cycle the master sc. you won't have to drop the domain but at least plan for it to. also, if the sc has been running in this faulted state for some time now and the domains are running as expected then i would push to get downtime (late night maintenance if need be). i dont think there is much more you can do at this point. do you have a service contract with sun for this box?

Thought I would give you all an update.

I stopped the various applications within the two domains, A and C Wanted to ensure no data was lost, and ensured everything was backed up.

So, I manually reset the sc0 controller via the reset button on the front. I

Domain A failed, error leds lite up on the baords, and was uncontactable. Domain C was ok, and I could log in etc etc.

Once the main SC controller had rebooted, I consoled onto Domain A, showkeyswitch was off, so I switched it on, 15 mins of diags as the boards came up. Then did I boot. Domain came backup again.

So, for anyone else that does a search, yes, its more than likely to crash a domain ! :-)

Cheers for your help guys.

SBK

You really should update the ScApp version to a later one if you can arrange a maintenance window.

yeap, will do. Difficult job to upgrade the firmware ?

sc i haven't done in a very long time but the xscf's are pretty easy.

this might help you on your way: http://dlc.sun.com/pdf/816-2972-10/816-2972-10.pdf