Originally posted by: SystemAdmin
Hi there,
Can someone help me to resolve the dacNONE error. Brief about env: AIX 5.2 TL10 SP1 with RDAC, smutil.aix.rte 8.40.650, 2 nodes cluster(HA 5.3), each node having 2 HBAs (fcs0 & fcs1) only for disks access with Concurrent VGs (Node A:RW, NodeB: Passive VG permission). Storage is FastT700 with EXP710 connected in the same rack. LUN's mapped to the hostgroup (node A and B). In working setup DAR0 has DAC0 and DAC1.
NodeB#> fget_config -vA
---dar0---
User array name = 'FAStT700'
dac0 ACTIVE dac1 ACTIVE
FCS1 in nodeB failed, errpt floods with fcs1 controller failure errors. LUNs preferred paths were moved to Controller A from B by FAStT and the affected disk paths have changed to FCS0 in both nodes. Hardware vendor confirmed fcs1 failure and replaced the failed fcs1 in nodeB by:
1. rmdev -l fcs1 -d -R (it removed dac1, fscsi1, fcnet1, fcs1)
2. replaced the hotplug HBA
3. run cfgmgr, (nodeB found fcs1,fcnet1,fscsi1)
4. diag on the new card with wrap plug successful
5. modified the alias on the SAN zone, saved and enabled the config
6. verified SAN path, cables OK
7. run cfgmgr, but still dac1 is missing in nodeB
User array name = 'FAStT700'
dac0 ACTIVE dacNONE ACTIVE
when I tried to move the LUN's back to their preferred controller B, some LUN's moved fine to the preferred controller, nodeA and nodeB disk paths also got changed. Just before completion of moving all the afftected LUNs, Fastt changed the moved LUN's path back to A from B.
I really dont know why FastT is moving the LUN's to the other controller. Something wrong with Fastt..? or the above procedures I followed to replace the HBA in concurrent VG env is wrong..? Can someone help me please.
Thanks in Advance,
Arul.
#AIX-Forum