VAPIC is Unavailable

[swelsch]

Hi,

i have a ACI mini and everything works fine. After updating ACI to 4.2(7f), one of the virtual apics gets Unavailable after 4-5 days and gets only available again after Reboot. After Reboot, the vApic i available for another 4-5 days before gets unavailable. The cluster shows fully fit. SSH the management IP to the vAPIC works, endpoints are learned, i can ping the endpoints from the VAPIC, everything looks perfect, but i can't ping the other apics 10.0.0.1 + 10.0.0.3

vaci02 and vaci03 are running on the same UCS, the same FI, connected to the same Leaf.

Has anyone an idea, what i can do to fix this problem ? is it a bug ?

thanks in advance

Steffen

[Robert Burns]

If things were stable in the previous software version (only started failing after the upgrade) I would suspect a bug of sorts.  Open a TAC case so they can confirm.  Typically when services fail after a relatively consistent period of time, could indicate a memory leak.  I would expect this behavior on all vAPICs though, so that doesn't necessarily fit here.

One thing to check is the DB integrity.  Look for any mutators in the output of "acidiag rvread" from each APIC.

Robert