Created attachment 815814 [details]
engine log

Description of problem:

Incident occurred after the rolling upgrade test from Big Bend to Big Bend U1 on one of the nodes while one of the nodes was rebooted with chkcfg off on glusterd and vdsmd. 

The status shown on the host in rhsc during reboot showed that the host went from non-operational back to unassigned and back to non-operational. Host should not have just remained in the non-operational state.

Description of rolling upgrade: 3 dist-rep volumes with 2 replicate pair count were created. One node from the replicate pair was upgraded and rebooted with glusterd and vdmd off.

Version-Release number of selected component (if applicable):
rhs: rolling upgrade from RHSS-2.1-20130910.0-RHS-x86_64-DVD1-BIGBEND.iso to u1
rhsc: big bend

How reproducible:


Steps to Reproduce:
1. create a 4x4 dist- rep volume
2. do a rolling upgrade on one of the nodes
3. reboot the server with chkcfg off on glusterd and vdsmd

Actual results:
host went to status non-operational
then it went to unassigned
back to non-operational


Expected results:
host should have remained in the non-operational state


Additional info:
vdsmd and engine logs attatched
went from non-operation to unassigned at 19:24:59
unassigned to up at 19:25:04