This morning, I went through the process of failing over. The load
balancers are behind a firewall. The firewall will take all external
traffic and redirect it on an internal ip to the load balancer. While I
was testing I pinged a farm every 3 seconds using its internal and
external ip.
LB-1 - Primary
LB-2 - Backup
1. Put LB-1 into maintenance mode
- This automatically makes LB-2 the primary
- This worked correctly. Both External and Internal pings ran
continuously.
2. Perform maintenance on LB-1.
3. Return LB-1 from maintenance mode
- This worked.
4. Force LB-2 into maintenance mode
- This should force LB-1 to be the primary
- This was the point of failure.
- LB 2 was put into maintenance mode, the firewalls said everything
happened correctly. Internal traffic ran continuously and did not drop
any pings, External traffic no longer went through. All pings were dropped
- When I goto the firewall, I cannot ping the internal ip address.
I'm assuming the load balancer would advertise the new MAC address of
the ip? Is it possible that on the switch back, it isn't correcting the
arp table?
thanks.
myles.
On 10/4/12 5:01 PM, Laura Garcia wrote:
The steps are correct. It will ne very useful a tcpdump between the
nodes while the swithes are running.
Regards
El 04/10/2012 21:54, "Myles Merrell" <[email protected]
<mailto:[email protected]>> escribió:
Laura,
I'm running the zen load balancers as kvm virtual servers on proxmox.
I was so concerned with everything not working, I don't remember
the specifics. The weird thing is that I tested this before I put
production traffic on it, and it was working correctly. I also
have a test setup with v3-rc1 on it, and doesn't experience the
problem. One thing I did do out of norm was to update the
password on the gui interface, but that's it.
I will do another test tomorrow morning, and take careful notes.
I will follow the instructions I listed below (assuming they are
correct). I will take a close look at each of the interfaces as a
I switch back and forth, and note exactly when things stop working.
Is there anything else I should be looking at during the process?
thanks.
myles.
On 10/4/12 3:46 PM, Laura Garcia wrote:
It isn't a normal behaviour. What is your configuration architecture?
Once the switch is happened, how does your network interfaces
appear at the active node?
El 04/10/2012 21:09, "Myles Merrell" <[email protected]
<mailto:[email protected]>> escribió:
I'm running r2-stable
There are https farms and http farms.
And all of the pings are lost.
Tomorrow morning at 5am EDT, I can run some more tests and
try different approaches. The problem is that it has some
production traffic on it.
thanks.
myles.
On 10/4/12 2:52 PM, Laura Garcia wrote:
Hi Myles, what version of zenlb are you running? Virtual or
physical? What kind of farms? How many pings are lost on
every node switch?
Regards
El 04/10/2012 20:01, "Myles Merrell"
<[email protected] <mailto:[email protected]>>
escribió:
I've tried a few times to perform some maintenance and
tuning tasks to
my clustered load balancers, and I've experienced downtime.
I was wondering if there is a basic set of instructions.
Something I'm
doing is causing the farms on each server to go down, or
become
inaccessible.
I have two clustered load balancers:
LB-1 - Primary
LB-2 - Backup
1. Put LB-1 into maintenance mode
- This automatically makes LB-2 the primary
2. Perform maintenance on LB-1.
3. Return LB-1 from maintenance mode
4. Force LB-2 into maintenance mode
- This should force LB-1 to be the primary
5. Perform maintenance on LB-2
6. Return LB-2 from maintenance mode
During the process my constant pings to farms, stop
working and people
cannot access the sites.
Is there a different process I should be using? Is
there something else
I should be looking at?
thanks.
myles.
------------------------------------------------------------------------------
Don't let slow site performance ruin your business.
Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java,
and .NET app
Try New Relic at no cost today and get our sweet Data
Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
--
Myles Merrell
Director of Software Engineering
[email protected] <mailto:[email protected]>
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy
New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and
.NET app
Try New Relic at no cost today and get our sweet Data Nerd
shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
--
Myles Merrell
Director of Software Engineering
[email protected] <mailto:[email protected]>
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New
Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
--
Myles Merrell
Director of Software Engineering
[email protected]
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
