What I would need:
Could someone with a RH5 cluster, with IPTables running, please post their 
iptables info? 
What I have:
I hate to sound like an idiot there, but I just can't understand what is 
happening with IPTables and linux HA.
I have 3 servers running RH5, dal-xcp-11(192.168.1.1), dal-xcp-12(192.168.1.2) 
and dal-xcp-13 (192.168.1.3) clustered together in an Active/Active/Standby 
configuration.
With IPTables turned off my cluster performs flawlessly.  (Finally.)
This is what my ha.cf looks like on all 3 machines:
    crm on
    auto_failback on
    logfacility     local0
    logfile /var/log/hb.log # Heartbeat logfile.
    debugfile /var/log/heartbeat-debug.log # Debugging logfile.
    apiauth  mgmtd   uid=root
    respawn  root    /usr/lib/heartbeat/mgmtd -v
    keepalive 10
    deadtime 30
    warntime 20
    initdead 120
    udpport 894
    respawn root /usr/lib/heartbeat/hbagent
    mcast eth0 237.0.1.1 894 1 0
    respawn hacluster /usr/lib/heartbeat/ipfail
    node dal-xcp-11.prodea-lo.net
    node dal-xcp-21.prodea-lo.net
    node dal-xcp-12.prodea-lo.net

The multicast IP and port I'm using are 237.0.1.1 port 894.  Along with the 
other lines in my normal IPTables setup, I add the following line for HA.  Am  
I missing something here?  I would have thought that adding these two lines for 
the IP and port would have been enough.  Does HB have other ports that it uses 
by default that I don't have listed?
    -A RH-Firewall-1-INPUT -d 237.0.1.1 -j ACCEPT
    -A RH-Firewall-1-INPUT -p udp -m udp --dport 894 -j ACCEPT
. . .
    -A RH-Firewall-1-INPUT -j LOG --log-prefix "Reject Traffic " --log-level 6
    -A RH-Firewall-1-INPUT -j REJECT --reject-with icmp-host-prohibited

Hell, I've even tried adding these lines, but it seems after a few minutes that 
it stops working too:
    -A RH-Firewall-1-INPUT -s 192.168.1.1 -j ACCEPT
    -A RH-Firewall-1-INPUT -s 192.168.1.3 -j ACCEPT
    -A RH-Firewall-1-INPUT -s 192.168.1.4 -j ACCEPT

When I turn on IPTables, I don't see any packets being rejected (and logged to 
/var/log/messages), but the multicast packets from the other two hosts stop 
showing up on the single server that I turned IPTables on for.
doign a tcpdump on port894 on the server that has IPTables turned on 
(dal-xcp-12), I see the following.  No packets from dal-xcp-11 or dal-xcp-21.  
On those two, I do see packets from all of the servers in the cluster.
15:06:11.116476 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221
15:06:21.117204 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221
15:06:31.117176 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
235
15:06:31.117194 IP dal-xcp-12.prodea-lo.net.32965 > 237..0.1.1.894: UDP, length 
221
15:06:41.118242 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221
15:06:51.118669 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221
15:07:01.119928 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221
15:07:11.121685 IP dal-xcp-12.prodea-lo.net.32965 > 237.0.1.1.894: UDP, length 
221

Quite often, at this point, one of the OTHER two blades reboots with the 
message:
dal-xcp-21 heartbeat: [3079]: EMERG: Rebooting system.  Reason: 
/usr/lib/heartbeat/crmd
But that (I think) is another issue
Any help would be greatly appreciated.
Michael.



_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to