Services Stays on the secondary server even the primary server comes back a

[zulbn]

Hi, I am researching whether it is possible if either service(ftp or ssl) goes down on the primary server, it will change the stickiness to the other secondary server. It should stick to this server even when the primary server goes back alive. The only way for the primary server to resume it's service is when the secondary server goes down.

Currently, I am using the PrimarySorryServer command but does not conform to their requirements. My customer wants it this way so that the critical data are correctly sync.

Anyone has such requirements. Thank you.

[zulbn]

Do I need to use script play to configure such requirements?

Your reply is greatly appreciated.

:)zul

[zulbn]

Can anyone help me on this? Cos I'm quite new to content switch configuration.

Rgds,

:)zul

[Gilles Dufour]

do you really need the command primarysorryserver.

For me I would see you just need to configure 2 services.

ServiceA and ServiceB

Add the services to your content rule.

And Configure stickyness - based our source ip for example.

Traffic will be loadbalanced between the services

but once a connection has been open with one service, all other connections from the same client will go to the same service, unless this service fail.

Usually, there is only one primary sorry server since we expect then when all other servers are down, this one should be up and contain an error message for the users to know services are down.

Regards,

Gilles.

[paul.matthews]

Keeping a user attached to an origin server is the whole point of stickiness.

Add both servers to the content rule and configure sticky (there are a few options on how you may want to do that) and an individual client will then stay with the server until the server fails.

Sorry servers are really just there for a "sorry, we ar having as problem at the moment, please try later" rather than a blank failure to get the web page.

Paul.

[zulbn]

Hi, thanks for the reply.

I used the primarysorryserver command as temporary solution. Cos what they need is that once "either ftp OR https" services fails, new connections will not use the primary server anymore (even when one service example ftp is up on the primary server). Any new connection will use the secondary server.

"Even when the primary server goes back alive", any new connections will not use the primary server but instead use the secondary server. Primary server will resume it's service only when I manually shutdown the secondary server. Only then, the primary server will resume the role. When secondary server goes back alive, new connections will still use the primary server.

Can it be done and how?

[Gilles Dufour]

if you are using a layer 3 content rule, the CSS will tie together ftp and https.

You could then use a keepalive script to check ftp and https. If one does not work, the service is removed.

So you will get something like

service ServerA

ip x.x.x.x1

keepalive type script ap-kal-check-ftp-https

service ServerB

ip x.x.x.x2

keepalive type script ap-kal-check-ftp-https

owner mycompany

content rulel3

vip x.x.x.x

add service serverA

add service serverB

advance-balanced sticky-srcip

active

!

There might be problem with FTP since this is a protocol using data channels.

But it may work if you are using the default port 21 for ftp.

Except that, there is no way to tie FTP and HTTPS rules and their sticky table.

Regards,

Gilles.