Question Regarding Load Balancing and Time Management
We have around 18,000 servers to be patched ever month we have around 2000 servers per shift which is 12 hours outage window. There are 9 shifts which covers all the servers, at present patches where per-staged and executed on outage windows. After staged patches execution We run around 5 jobs in batch to make sure everything are patched with zero missing patches at the end ( Reboot with NSH, Analyse and Patch , Analyse only and Last Reboot check ) .
These 2000 server per shift are spread across the static group which each has 50 server max means 2000 server are divided by 50 so 40 static groups are created. And these Batch jobs are executed against these 40 static group in parallel by four person like each person will have one batch job for one static group so there will 40 batch job running in parallel.
But the thread count of individual job in the batch job is only 50 in parallel, simply job parallel thread count alines with static group server count. We dont have any issues with present procedure with load balancing and time management.
We are planning to move to smart group to sync with change management portal for loading the servers based on the change outage window , so the real concern now is all these 2000 server will be loaded into the single smart group and jobs parallel thread count is only 50 at present. The question is to raise the parallel thread count of the jobs to some higher limit (Unlimited or 2000 ) cos if we run only 50 parallel thread its going to double the time so we dont want that to happens. By the same time we dont want hamper the BL infrastructure with high load.
I know that we are using the same high load in the present procedure which we use at preset without no issues , what is optimistic way to run with single smart group of 2000 server and also what is optimistic parallel thread count for each job
We have 4 app server and 10 Jobs servers to manage these loads with BSA 8.3 very soon we will upgrade into 8.6 SP1