Hi -
I have Samba 3.2.3 running on two Linux machines,
and there is Linux HA (v1) set up on these machines(active-passive).
This is what I have in my Heartbeat config file :
logfacility daemon
node MyNode1
node MyNode2
keepalive 1
deadtime 10
bcast eth0
ping my.ping.host.addr
auto_failback no
respawn hacluster /usr/lib64/heartbeat/ipfail
My smb.conf:
[global]
oplocks = no
msdfs root = yes
guest account = myaccount
workgroup = MYWORKGROUP
security = SHARE
max smbd processes = 15
interfaces = eth* lo
bind interfaces only = yes
[MyShare]
path = /opt/myshare
read only = No
guest ok = Yes
locking = no
hosts allow = valid.hosts.ip.addr
hosts deny = ALL
When I test a failover from one node to another, a Samba client
(either Linux or Windows)
does't reconnect automatically to the new node.
I was able to fix this adding this line to ha.cf:
"respawn root /usr/sbin/rcsmb" -- restarting Samba server
Is that a correct way of handling the issue?
Why doesn't a Samba client reconnect automatically after the HA
failover?
Thanks,
Sergey
|