MQSeries.net :: View topic

sm138929 · Posted: Mon Mar 31, 2008 9:53 am Post subject: Solution?

I did download the MS03 Supp pack ..it is indeed a novel idea to have the Qmgr objects recorded ...

Regarding backup then what will be the overall strategy ..I mean to say what is the recommended way to have backups of MQ servers in production so that even we don't have a DR site we can restore it from tape library ...Also we do understand that we are losing some data/logs because the backup has to be taken at a prior time ..
Also QMGRs inDC1 and DC2 are identical in all respects except the fact that the channel def in SDR chl has to be changed because the CONNAME parameter is having a IP specific to the coresponding datcentre Load balancer v-ip
.One question I have we have Sun cluster managing the 3 qmgrs QM0,Qm1 and QM2 in DC1 which are running in 2 solaris boxes with Qm0 as full repository in a MQ cluster CL..We have the same def of the MQ cluster in DC2 but there we have only one Solaris box with all 3 qmgrs running (with no sun cluster )....rest of the objects are exactly that of DC1 since we configured the DC2 qmgrs by using the SRDF copy to the DC2 SAN..Also whe nwe fail back toDC1 our plan has to change the SDR chl def>conname to the Ipof the Load balncer V-ip specific to DC1`

In our failover exercise.we have plan to take backup of the mqm data and logs as well as use the suup pak to take snapshot of the QMgr objects in DC1 because that is the production environment ..which we cannot risk...

Currently SRDF is stopped between DC1(prod) site and DC2(failover site ).Before we start the file copy frm DC1 to DC2 we will ensure that all messages are processed in DC1 apps and wil have the Sender andreceiver chls with our vendors stopped so that no new msgs arrive in our queues andhave the client chls stopped too ...We will also ensure that the queues in DC2 are completely cleared beforewe transfer the data using SRDF ...
During failback we are little bit apprehensive if the env we have after the DC2 data copies to production DC1 we have all the QMGRS functioning properly and chls getting live ,,We know that we nee to do reset--resolve --start .....

Regarding the failover steps do you recommend that we do the entire stuff in a stage/preprod MQ servers which we do have in DC1 datacenter and configure the DC2 MQ server as stage .

Thanks..

SM