|
Cristi wrote:
> Joseph Mack NA3T wrote:
>
>> On Mon, 9 Jul 2007, Cristi wrote:
>>
>>
>>
>>> Joseph Mack NA3T wrote:
>>>
>>>
>>>> On Sun, 8 Jul 2007, Cristi wrote:
>>>>
>>>>
>>>>
>>>>
>>>>> Hello again,
>>>>>
>>>>> I installed heartbeat an ran into some trouble. I run 2 services on the
>>>>> 2 real servers : httpd using LVS-NAT and sql using DR.
>>>>>
>>>>>
>>>>>
>>>> did you handle the different default gw requirements for
>>>> the two forwarders?
>>>>
>>>>
>>>>
>>> For my test setup, the real servers had as def gw the IP of director2
>>> because the VIP was active on it.
>>>
>>>
>> so you haven't handled it. Go look in the HOWTO and just try
>> one of LVS-NAT or LVS-DR at a time
>>
>> Joe
>>
>>
>>
>>
> For troubleshooting purposes, i had shut down director 1.
> The def gw for the real servers is the DIP on director 2, who's also
> masquerading for them.
> The DR part is working for ~12hrs now, so no need to test any further.
>
> Thank you for helping,
>
> Cristi
>
> _______________________________________________
> LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>
Problem solved.
I had fixed the hearbeat issue, meaning that i had reconfigrured HB to
do the LVS-NAT correctly without me having to do it "by hand" or script :).
I had fixed the default route issue also, but the first problem (meaning
the IP wouldn't reply to icmp from time to time) was still active. I had
narrowed it down to some arp issue and upon receiving an email from my
ISP i saw that the MAC address in the ARP entry for the VIP wasn't on my
directors. Upor further ISP investigation i discovered that someone in
the same broadcast domain was using that IP, making it unreachable from
outside. Once that problem solved, my troubles went away. I want to
thank Joe and Graeme for their help with this issue.
|
