Domino server mails not opening in the event of failover

[mayu]

Hi all,

We are using Domino server 9.0 and EV 10.0.4 mailbox and journal archiving

We have added one of our primary domino server in the EV target container and everything is working fine.

But whenever our domino server has failover to DR Domino servers users are not able to open their archived emails also journal archiving failed.

We have not added DR Domino mailbox servers in the EV target container yet. In order to open the emails from DR Domino server what should be the next action plan.

Do we need to add DR Domino servers in the EV target container and all prerequisites need to be follow? Please suggest.

[EV_Ajay]

Hi,

I think you need to add EV Service account and provide all permission on DR Domino Server. Please check and compare any other things between current Domino server and DR Domino Server.

[MassimoVagli]

Hello, 

I guess you have a domino server cluster, in that case you need to add the second Domino server (DR) as target for Enterprise Vault archiving and follow the wizard for the domino server cluster acknowledgement (basically you need to check one of the two servers like the 'preferred' one).

In addition after the failover I guess you need to re-execute the provisioning task, in that case Enterprise Vault will 'understand' that the dominoserver1 is no longer available.

Let me know how it goes.

Massimo

[Arjun_Shelke]

I agree with Massimo, you will need to add DR Domino servers as Targets in Enterprise Vault.

[Paul_Honey · Accepted solution]

Hi,

Just a little more explanation on the responses above. All of EV's API calls to the Domino environment are cluster aware (i.e. when trying to access a database, we use the OpenWithFailover method) but there are other dependencies which I will try and explain below

1. For journal archiving

Journal databases (i.e. mailjrn.nsf) are not recommended to be replicas in a Domino clustered environment and by default will not be anyway. The Domino server creates its own non-replica journal NSFs when the router task starts and manages those NSFs local to the server according to the rules set on the server's corresponding configuration document.

When your primary server is up, I expect the majority of mail routing, and consequently journaling, to be handled by your primary server, so EV needs that server as a target with the corresponding journal location configured. However in times of failover, either because the primary server is down or overloaded, then the failover server will take on mail routing, and hence also journaling, responsibilities. As a result, that server also needs to be added as a target to EV, and its journaling location configured, to ensure we ingest from that server's journal nsf's as well

2. For mailbox archiving

As in a Domino clustered environment, mail files will be replicated on both primary and failover servers, if the primary server is down or overloaded, then the mailbox archiving task will automatically failover to the replica on the failover server. As such, adding the secondary server as a target for EV is not a 100% pre-req in a mailbox archiving only environment, although adding it causes no harm / disadvantages

In fact, adding the failover as a target server opens up an interesting configuration opportunity in an Active-Passive Domino clustered environment (i.e. all users home mail server is set as the primary server) that many of our customers use. We provide the configuration option called 'Preferred Server', as mentioned by Massimo, for Domino clustered servers (on the Domino server target properties) which would allow you to set the failover server as the preferred target for the mailbox archiving task. This is useful if your primary / home mail server is overworked, as you effectively can tell the mailbox archiving task to prefer to do its work against the failover server, archiving data and creating shortcuts on that server, which will subsequently be replicated back through normal cluster replication to the primary mail server.

3. For mailbox archiving retrieval

When you say users cannot retrieve data in a failover situation, I assume this is because your EVDG server is also 'down'? We unfortunately do not support Domino clustering of EVDGs, as the EV retrieval architecture does not easily lend itself to that. So, if the EVDG is down, and you have failed over to a DR server that is also configured to act as an EVDG, then as Massimo states, you need to re-run the provisioning task to push out the name of this 'new' EVDG to all mail files / users to use

As the EVDG is actually a component / extension of an EV server, you can provide more automated resilience / HA / DR by clustering your EV Server using VCS or MSCS, and including the EVDG Domino server as a clustered application / service to allow the same Domino server to run on primary or failover nodes depending on the MSCS / VCS cluster controls

I hope that helps you understand the various components and influcences of clussetering in an EV for Domino environment further

Regards

Paul