Service Desk Practitioners Forum
cancel

Integration of Omnibus with a resilient Service Desk set-up

Highlighted
CA1457266
Respected Contributor.

Integration of Omnibus with a resilient Service Desk set-up

Hi All,

We need to integrate Omnibus with SD application server. But at the same time, we need to have a resilient structure for Service Desk.

I need someone to help me in finding whether the below mentioned set-up is possible:-

we will be using two SD application servers. CSS is being used to obtain resiliency for these two servers.The IP address of these two application servers is mapped to a common Virtual IP address. All the request come at this Virtual IP address.At a time only one SD application server is active (SD1) and the other (SD2) is in warm stand-by mode. So all the request are routed to SD1 application server. If the SD1 application server fails at any point of time, then failover will be obtained by switching over to the SD2 application server.

1) If such a set-up is poissible then can Omnibus still raise incidents when the failover happens?
2) Doesnt this require any kind of setting to be done on the Omnibus side? or the failover from SD1 to SD2 wont have any effect on the procedure of Omnibus raising the tickets in SD?

Any info will be useful.

Thanks in advance.

Regards,
Anu.
3 REPLIES
Mike Bush
Acclaimed Contributor.

Re: Integration of Omnibus with a resilient Service Desk set-up

It sounds as if Omnibus is using SD_Event or Data Exchange to connect to the HP SD service, as such it will use the Virtual IP address and should fail over when the servers switch. But you are subverting the inbuilt load-balancing mechanisms of HP SD and what happens when you "switch" - I suspect the clients are thrown off and need to login again - whereas with both app servers "live" and using HP "multiple servers" mode a failure is seamless to the client user (you may have other reasons for only wanting one server live though?)
CA1457266
Respected Contributor.

Re: Integration of Omnibus with a resilient Service Desk set-up


Thanks for the information Mike.

So if we have the "multiple server" mode on, with both the servers live, will the virtual IP work in this scenario?

So you mean to say

1) All the request will be routed to the virtual IP address and then it will be forwarded to one of the live servers?

2) Will both the servers resolve the client request or only one (primary application server) will be resolving the client request?

3) Also, when the primary server goes down the inbuilt load-balancing mechanism will switch to the secondary application server and so now the secondary application server will be resolving all the request?

4) Also with all this set-up in place can Omnibus still raise incidents when the failover happens?

Any info would be useful.

Regards,
Anu
Mike Bush
Acclaimed Contributor.

Re: Integration of Omnibus with a resilient Service Desk set-up

I think the virtual IP will route to either of the App. Servers in the first instance (I normally advise a DNS entry with "round-robin", is this similar?) Once connected a file called SRV.DAT is set up on the client machine with the IP address of ALL the live App. Servers. Then the "load balancing" within HPSD reserves the right to direct clients to an alternate IP address (assuming these are visible to clients in your setup and that there is no "firewall" effect), similarly failure to connect gives the client a list of alternate IP addresses to try. I suggest you set it up and see what gets into SRV.DAT on a client machine. The SD_EVENT / DATA_EXCHANGE connections won't load balance or fail over in the same way but should be able to benefit from your virtual IP.