VPAR : second vpar reboot causes first vpar rebooting.

[Arun B S Nair]

Hi Gurus,
I have a rp8420 server having one npar running two vpars. ( HPUX 11.11 , Vpar A.03.03.06 , PDC 23.1 , STM 47.15 )
This npar is having two single core cells and a total 8 cpus. Each vpar configured with 4 cpu each( one cell board each ). A reboot of Vpar1 is not causing any problem to Vpar2.But when I reboot vpar2 , it resets vpar1 also.
( no updates in shutdown log or syslog of vpar1 for this ). In VFP, during this reboot I can see the partition status as "halted" for a while. Any thoughts ... Many Thanks in advance

[IT_2007]

can you also post vparstatus output without -v options.

[Arun B S Nair]

here is the o/p
aded119p:root [/stand] vparstatus
[Virtual Partition]
Boot
Virtual Partition Name State Attributes Kernel Path Opts
============================== ===== ============ ======================= =====
aded119p Up Dyn,Auto,Nsr /stand/vmunix
aded115n Up Dyn,Auto,Nsr /stand/vmunix

[Virtual Partition Resource Summary]
CPU Num Memory (MB)
CPU Bound/ IO # Ranges/
Virtual Partition Name Min/Max Unbound devs Total MB Total MB
============================== ================ ==== ====================
aded119p 4/ 6 4 0 11 0/ 0 24576
aded115n 2/ 4 2 2 11 0/ 0 24576

[Tim Nelson]

SO you are simply executing shutdown -r -y now on aded119p and nothing else ?

No , no RS ?

I have a couple similarly configured rp7420's with no issues during reboots from either server.

[IT_2007]

Looks good and no problem in output. Only thing is if you don't see any errors in syslog on both vpars, my suggestion is look at GSP chasis log. It may give you some idea. If you still stuck at that point, time to call HP.

[Arun B S Nair]

Yes I am issuing only shutdown -r -y 0

[IT_2007]

also would you please check size and permissions for /stand/vpdb and vpmon files on both servers? Please post results.

[Joël Techer]

Hi,

I suggest you to do vparextract -l command that will be very interesting for you => VPMON log
so you will be able to detect the pb

Has the vpars newly created?

Best regards

Jean-Louis

[Arun B S Nair]

both servers are having vpdb and vpmon of same size. vparextract -l is not giving anymore insight to the issue.

Another observation is that,
a shutdown -h -y 0 of any of these partitions funcions properly and doesn't reboot the other one ..

[Mridul Shrivastava]

I would suggest you to check the firmware level, there have been issues if the firmware doesn't match the minimum vpar version criteria:

Please refer the following link for the same ( look at page 12):

http://docs.hp.com/en/1705/oc.pdf

[Arun B S Nair]

I could resolve the issue by reconfiguring the vpars. My understanding about the issue is that,
Cell 0/10 was the monarch cpu for the npar.( please refer to my earlier attachment ) Since it was a bound cpu for vpar aded119p and that was holding the least
LUN ID out of all cells, the vpar also took it as its monarch cpu. When that vpar reboots this cpu get resets, causing the Npar to be restarted.

While reconfiguring the vpars, I created vpars without having this cpu as a bound CPU. ( it is configured as a floating one currently). Further rebooting didn't give any problems.
I am attaching latest vparstatus with this.
Please post your thoughts about this issue. We are consulting HP to know if this is a known issue.. Any documents related to this will be highly appreciated..