cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
618
Views
0
Helpful
1
Replies

server in "discovery failed" after a FI reboot

vincent paulin
Level 1
Level 1

Hi all,

Today  did acceptance tests on 2 UCS mini  (flashstack configuration). 
On both configuration I got the same problem.
To test the redondancy of FC and uplinks paths on the UCS I rebooted a FI .
As I wanted to validate the redondancy I first did the problem on the subordinate FI, then after everything came back OK I did the test on the other FI.
ON both UCs I got that problem:

When the first FI I rebooted (FI-A for one chassis, FI-B for the other one) the server in the last occupied slot (5 in both cases) got to error type "Discovery failed".
The service profile (an ESXi) on that server was still running but only with FC and uplinks from the peer FI.

I had to rediscover that server to fix the problem.

During all  other tests I did after  (rebooting the same FI,  the other one, or even fully power off/On the UCS mini to validate wmware-ha to the other site) i never encontered that problem.

As the problem was similar on 2 new chassis , maybe it 's a bug.
The UCS installed version is 3.2.(3i) .

Do you ever heard about a same problem ? Got an idea of what could happend ?

Thanks in advance

1 Reply 1

cassiolange
Level 1
Level 1

 Hello Vincent, 

 

Without any log, is difficult see why this happened.  

 

You could do this test again?  Relocate a blade from another slot to slot 5 and perform the test again. 

 

IF the fail move together with the blade,

  1. Investigate this blade. 

If fails again on slot 5, two options

  1. open a tac case 
  2. update the UCS  environment

Regards and good luck,